Evidentiality and Epistemic Modality: Conceptual and Descriptive Issues 9783034339933, 9783034346511, 9783034346528

Table of contents :
Cover
Series Information
Copyright Information
Table of Contents
Introduction
Acknowledgments
Section A Evidentiality: Conceptual Issues
On Inferential Evidentiality: Is “Evidential” Inference Abductive? (Patrick Dendale, Johanna Miecznikowski)
Further Motivating the Challengeability Criterion for Evidentiality (Tabea Reiner)
Section B Evidentiality and Modality: Descriptive Issues and Corpus-based Studies
Evidential Strategies and Hierarchies in Ladakhi: the Case of Sensory Perceptions (Nicolas Tournadre)
Inferential must from a Contrastive Dutch-German Perspective: Semantics and Constructional Aspects (Tanja Mortelmans)
Conjectural Future in French and in Spanish: An L2 Acquisition Perspective (Aoife Ahern, José Amenós-Pons, Pedro Guijarro-Fuentes)
Romance Future: between Epistemic Modality and Evidentiality (Cecilia Mihaela Popescu, Oana Adriana DuȚă)
Speaker Commitment in Reportative and Folklore Evidentiality: The Case of Spanish ‘o eso dicen’ (Dorota Kotwica)
Toward an (Exclusively) Inferential Marker: The Modal deure (‘mustʼ) in Contemporary Catalan (Andreu Sentí)
Evidentiality and the Latvian Oblique Forms (Andra Kalnača, Ilze Lokmane)
Section C Evidentiality and Modality in Discourses and Genres
Evidential Expressions in Spanish Accounts of Religious Miracles of the 17th Century (Natalia Mora-López)
The Use of the English Modal Verbs in Linguistic and Philosophical Research Articles (Liisa Vilkki)
Notes on Contributors
Series Index

Citation preview

Evidentiality and Epistemic Modality

Linguistic Insights Studies in Language and Communication Edited by Maurizio Gotti, Emeritus Professor, University of Bergamo (Italy). Volume 297

ADVISORY BOARD Vijay Bhatia (Hong Kong) David Crystal (Bangor) Konrad Ehlich (Berlin / München) Jan Engberg (Aarhus) Norman Fairclough (Lancaster) John Flowerdew (Hong Kong) Ken Hyland (East Anglia) Roger Lass (Cape Town) Françoise Salager-Meyer (Mérida, Venezuela) Srikant Sarangi (Cardiff) Susan Šarcˇevi´c (Rijeka) Lawrence Solan (New York)

PETER LANG

Bern • Berlin • Bruxelles • New York • Oxford

Marta Carretero, Juana I. Marín-Arrese, Elena Domínguez Romero, Mª Victoria Martín de la Rosa (eds.)

Evidentiality and Epistemic Modality Conceptual and Descriptive Issues

PETER LANG

Bern • LANG Berlin • Bruxelles • New York • Oxford PETER

New York • Berlin • Brussels • Lausanne • Oxford

Bibliographic information published by die Deutsche Nationalbibliothek Die Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available on the Internet at ‹http://dnb.d-nb.de›. Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Research Project FFI2015-65474-P: Evidencialidad: Estudio Discursivo-Pragmático del Inglés y otras Lenguas Europeas. Evidentiality: A Discourse-Pragmatic Study of English and Other European Languages (EVIDISPRAG). Ministerio de Economía y Competitividad & Fondo Europeo de Desarrollo Regional (MINECO/FEDER).

ISSN 1424-8689 E-ISBN 978-3-0343-4651-1 (E-PDF) DOI 10.3726/b20323

ISBN 978-3-0343-3993-3 (Print) E-ISBN 978-3-0343-4652-8 (EPUB)

© Peter Lang Group AG, International Academic Publishers, Bern 2023 [email protected], www.peterlang.com All rights reserved. All parts of this publication are protected by copyright. Any utilisation outside the strict limits of the copyright law, without the permission of the publisher, is forbidden and liable to prosecution. This applies in particular to reproductions, translations, microfilming, and storage and processing in electronic retrieval systems.

Table of Contents

Introduction ������������������������������������������������������������������������  7 Acknowledgments �������������������������������������������������������������� 13 Section A  Evidentiality: Conceptual Issues Patrick Dendale and Johanna Miecznikowski On Inferential Evidentiality: Is “Evidential” Inference Abductive? �� 17 Tabea Reiner Further Motivating the Challengeability Criterion for Evidentiality ��������������������������������������������������������������������������������  73

Section B Evidentiality and Modality: Descriptive Issues and Corpus-based Studies Nicolas Tournadre Evidential Strategies and Hierarchies in Ladakhi: the Case of Sensory Perceptions ������������������������������������������������������������������������� 113 Tanja Mortelmans Inferential must from a Contrastive Dutch-​German Perspective: Semantics and Constructional Aspects ������������������������������������������� 137 Aoife Ahern, José Amenós-​Pons and Pedro Guijarro-​Fuentes Conjectural Future in French and in Spanish: An L2 Acquisition Perspective ��������������������������������������������������������������������������������������� 173

6

Table of Contents

Cecilia Mihaela Popescu and Oana Adriana Du Ță Romance Future: between Epistemic Modality and Evidentiality ������������������������������������������������������������������������������ 203 Dorota Kotwica Speaker Commitment in Reportative and Folklore Evidentiality: The Case of Spanish ‘o eso dicen’ ���������������������������� 227 Andreu Sentí Toward an (Exclusively) Inferential Marker: The Modal deure (‘mustʼ) in Contemporary Catalan ��������������������������������������������������� 251 Andra Kalnača and Ilze Lokmane Evidentiality and the Latvian Oblique Forms ���������������������������������� 281

Section C Evidentiality and Modality in Discourses and Genres Natalia Mora-​López Evidential Expressions in Spanish Accounts of Religious Miracles of the 17th Century ����������������������������������������������������������� 319 Liisa Vilkki The Use of the English Modal Verbs in Linguistic and Philosophical Research Articles ������������������������������������������������������ 347 Notes on Contributors ���������������������������������������������������������������������� 383

Introduction

This volume brings together cutting-​edge research on evidentiality and epistemic modality, covering conceptual issues as well as descriptive studies on a number of languages, discourses and genres. The contributions are mostly revised and enlarged versions of papers presented at the International Conference on Evidentiality and Modality 2018 (ICEM’18), held at Universidad Complutense of Madrid on 16–​19 September 2018. This conference provided an exceptional opportunity for academics from many countries to share their research and exchange their views; the resulting insights and approaches are partly reflected in the present book. The volume is divided into three sections, which correspond to three main themes. Section A contains two papers focused on theoretical issues on evidentiality, namely the nature of evidential inference and the criterion of challengeability for setting apart epistemic modality and evidentiality. Section B covers descriptive work, mostly based on corpora, on a number of languages: English, Spanish, French, German, Dutch, Italian, Romanian, Catalan, Latvian and Ladakhi (a Tibetic language); some of these papers set forth contrastive analyses of two or more of the languages listed above. Section C contains two papers focused on specific discourse domains and genres, namely accounts of religious miracles and linguistic and philosophical research articles. Section A starts with Chapter 1, by Patrick Dendale and Johanna Miecznikowski, which explores the nature of the inferential process involved in inferential evidentiality. This topic has been extensively treated in the French tradition through the work by Jean-​ Pierre Desclés and Zlatka Guentchéva, who have claimed that evidential inference is fundamentally abductive in nature. Drawing on more than 40 publications by these authors, this paper presents and assesses their abduction claims called “narrow claim” and “broad claim”, and proposes that the “narrow claim” is the useful one for the analysis of

8

Introduction

inferential evidentiality. The paper describes Desclès and Guentchéva’s abductive inference schema, which comprises three elements: an observation, a general law and a hypothesis concluded from these premises. Finally, Dendale and Miecznikowski point out a series of issues and unanswered questions raised by Desclès and Guentchéva’s analysis and characterization of abduction. Chapter 2, by Tabea Reiner, questions the status of challenge­ ability as a criterion to set apart epistemic modality from evidentiality, in the sense that epistemic meanings can be challenged while evidential meanings cannot, and then denies this status on the grounds that, in contrast to temporal, modal and aspectual meanings, evidential meanings do not operate on claims, although they may constitute claims on their own. Reiner also sketches a framework for evidentiality which distinguishes two levels, one not drawing on claims and the other consisting of claims. These levels are related to the grammatical or lexical status of their markers. Section B, covering descriptive work on a number of languages, starts with Chapter 3, authored by Nicolas Tournadre, which concen­ trates on the marking of sensory perceptions and evidential strategies and hierarchies in Ladakhi (a.k.a. Ladaks), a language that displays a rich evidential-​epistemic system in accordance with its membership in the Tibetic language family. The paper shows that the sensory markers, DUG and RAG, which play an essential role in Central Ladaks, cannot be simply described as “visual” and “non-​visual”. Instead, various factors, such as the hierarchy of evidential functions, temporal scope and shareable sensory experience, play an important role in the choice of evidential markers. Through the analysis of a parallel corpus of detective novels, Chapter 4, by Tanja Mortelmans, sets forth a detailed account of the reasons why English epistemic-​inferential must is much more common than its counterparts moeten (Dutch) and müssen (German) with this meaning. English must is found to have a wider epistemic range in a number of respects: it is compatible with weaker epistemic judgments; it does not only code inferences on the basis of external evidence or world knowledge but also expresses speaker-​oriented conjectures that lack a clear evidential basis; it is used to express empathy with the addressee, while Dutch and German seem to prefer other options for

Introduction

9

this purpose. This different behavior of must can be related to its stronger grammaticalization compared to its Dutch and German cognates, which are still prominently geared toward expressing (less grammaticalized) deontic and dynamic modality. The paper also proves translators to highly capture the subtle linguistic differences under study. Chapters 5–​ 8 cover issues on evidentiality and/​ or epistemic modality in Romance languages. Chapter 5, by Aoife Ahern, José Amenós-​Pons and Pedro Guijarro-​Fuentes, examines the interpretation of the conjectural and concessive future in Spanish by non-​native L1 French speakers at upper-​intermediate and advanced levels. An empirical study was carried out, based on acceptability judgments of clauses containing verbs with different lexical aspects and comparing the responses of the French speakers with those of a control group of native speakers of European Spanish. The results show that the French speakers fully transferred the L1 features; this transfer had a hindering effect regarding the conjectural and concessive uses of the simple future but led to a facilitative effect in the use of the compound future. The results do not differ significantly across the two proficiency levels, which suggests that the feature reassembly required for full command of the simple future in Spanish poses difficulties for L1 French speakers. The research provides no definite answer as to whether these speakers will eventually acquire the uses of the Spanish future that do not correspond to the French future; however, they showed sensitivity to the [± telic] feature, which suggests some positive evolution and opens the possibility for subsequent progress. The future tense in Romance languages is also approached in Chapter 6 by Cecilia-​Mihaela Popescu and Oana-​Adriana Duţă, which focuses on the so-​called “epistemic future” in adversative and concessive contexts in Italian, Spanish and Romanian. The chapter proves that a complex rhetorical strategy is involved in such contexts, where the epistemic future actualizes the prototypical feature of placement in a subsequent relation, exclusively marking the relation of subsequence of the hypothesis compared to a previously mentioned state of facts. In its first part, the paper sets forth a comparative analysis of the semantic and functional behavior of adversative and concessive uses of the epistemic future; this analysis mainly emphasizes that a description of inferential cognitive processes is actualized in both the adversative

10

Introduction

and the concessive contexts. The second part shows that the epistemic future in these contexts does not have an evidential reportative value per se; instead, this value is a meaning by default, which has to be actualized through the semantic and syntactic framework in which the verb morphemes occur. Chapter 7, by Dorota Kotwica, analyzes the evidential mean­ ing of the Spanish construction o eso dicen (‘or so [they] say’) from the perspective of speaker commitment and involvement. This construction can be used either to indicate that the evidence originates from a roughly defined group of people (reportative) or to introduce pieces of common knowledge, such as popular sayings and proverbs (folklore). The study contains a theoretical discussion of these two subtypes of evidential meaning, which shows differences in the degree of speaker commitment. In reportative contexts, additional elements of evaluation often stress a speaker’s dissociation from the source. However, in folklore contexts, speakers rarely show doubt regarding the information or its source, since folklore evidence is based on universally shared knowledge within the community to which the speaker belongs; this means that, even if speakers have no sensorial mode of knowing, part of the speakers’ perspectives are reflected in their utterances. The chapter also offers an empirical study of the use of o eso dicen with reportative and folklore evidentiality, based on a sample from El Corpus del Español by Mark Davies. Chapter 8, by Andreu Sentí, sets forth a study of the modal construction with the verb deure (‘must’) at the beginning of the Contemporary Catalan period (from 1833 onwards), based on a number of written and spoken corpora. The results show that, in the 19th century, the auxiliary deure consolidates the inferential evidential value, which had already emerged in Old Catalan, while its original deontic modality value decreases. During the 20th century, deure loses its deontic value in most Catalan dialects, thus differing from its correlates in the other Romance languages. The chapter also discusses new subjective values of deure (generic inference and conjecture), as well as two pragmatic intersubjective uses, namely a value of courtesy and another of inference about a future state of affairs. The paper also points out that, while in previous works on Old Catalan the inferential reading of deure had been considered evidential and not epistemic, the subjective and

Introduction

11

intersubjective uses found in data from the 19th to 20th centuries challenge the relation between evidentiality and epistemic modality. Section B ends with Chapter 9, by Andra Kalnača and Ilze Lokmane, which examines the evidential meanings of the oblique forms in Latvian and their interaction with other meanings. The study, based on a corpus and other authentic data, shows that the main meaning of the Latvian oblique forms is reportative (either quotative or hearsay), thus being evidential. This evidential meaning remains in all oblique uses, regardless of whether the discourse does or does not mention the source of reported information. By contrast, epistemic and mirative meanings are superimposed overtones, largely dependent on pragmatic factors, and most manifest in the presence of additional modality indicators such as modal particles or special syntactic constructions. The oblique may also express evidentiality combined with deontic modality, in the cases of reported imperative and reported necessity. Such a combination takes place mainly in complex sentences with verba dicendi in the main clause, the deontic meaning referring to the reported proposition. Section C, which comprises two papers on specific genres, starts with Chapter 10, by Natalia Mora-López, which analyzes evidential expressions in three 17th-​century accounts of religious miracles. The texts display a wealth of devices of reported evidence as well as verbs of seeing and hearing to support their narratives: people say something (typically quoted), people see something (e.g. miracles and signs), and people hear something (e.g. saints’ voices). This wealth of evidential expressions is not surprising, given the key role of veracity in this text type: even if accounts of religious miracles were commonly reported and widely accepted in the Spain of the 16th and 17th centuries, the narrated events were still perceived to look unbelievable, unnatural or even supernatural, hence the need for support by evidence in order to convince readers of the truthfulness of the accounts. Finally, Chapter 11, by Liisa Vilkki, sets forth a comparative analysis of the use of English epistemic modals (could, may, and might) and inferential modals (must and should) in research articles extracted from five linguistic and five philosophical internet journals. The chapter mainly draws on the metadiscourse framework and has also considered other studies on academic discourse and politeness. The results

12

Introduction

show that these modals often have similar, multifunctional uses in both types of articles. However, important differences were also found: most remarkably, philosophical articles include a higher amount of various predominantly reader-​oriented usages of these modal verbs. This difference may be explained by the nature of philosophical research, which mainly involves re-​assessing prior positions and presumptions. This difference seems to suggest that philosophers should take special care to control the level of the personal dimension in order to build convincing arguments. In summary, the chapters in this book are instances of ground-​ breaking research in evidentiality and epistemic modality, based on authentic linguistic data and including discourse, semantic or pragmatic perspectives. The two chapters in Section A provide thought-​ provoking views on two burning issues in present-​day research on evidentiality, namely the nature of the inferential process involved in inferential evidentiality and the role of challengeability as a criterion for setting apart epistemic modality from evidentiality. The descriptive studies in Section B are in-​depth studies of the way in which a number of linguistic devices express evidentiality and modality in specific languages. The two papers in Section C show how evidentiality and/​or modality are used by writers of different genres as devices for orienting communication toward the fulfillment of their aims.

newgenprepdf

Acknowledgments

As volume editors, we would like to acknowledge support from a number of research agencies who helped fund the participation of different researchers in the ICEM’18 Conference and the financial support provided by the following research project to enable the publication of this volume: •

Research Project FFI2015-​65474-​P: Evidencialidad: Estudio Discursivo-​Pragmático del Inglés y otras Lenguas Europeas. Evidentiality: A Discourse-​ Pragmatic Study of English and Other European Languages (EVIDISPRAG). Ministerio de Economía y Competitividad and Fondo Europeo de Desarrollo Regional (MINECO/​FEDER)

Section A Evidentiality: Conceptual Issues

Patrick Dendale and Johanna Miecznikowski

On Inferential Evidentiality: Is “Evidential” Inference Abductive?

Abstract Evidential markers of the type “Inferring” (Willett 1988) present the information in a sentence as the conclusion of an inference carried out by the speaker. The nature of this inferential process has hardly attracted the attention of scholars working in the field of evidentiality, especially in the Anglo-​Saxon literature. Instead, the focus was laid on the nature of the evidence an inference is based on, proposing distinctions such as sensory or perceptual evidence vs. non-​sensory or non-​perceptual, conceptual evidence, which, for example, underlie the distinction Inferred vs. Assumed made by Aikhenvald (2004). In the French tradition, how­ ever, the nature of the inferential process was explicitly described by two scholars, Jean-​Pierre Desclés and Zlatka Guentchéva, who have claimed since the mid 1980s that evidential inference is fundamentally abductive in nature (as opposed to deductive, for instance). Drawing on more than 40 publications by these authors, we here present their claim in detail and briefly assess it. In future work, we plan to review the way inferential evidentials have been characterized in the wider field of evidential typology and to suggest a new, more general, way of characterizing evidential inference, which will integrate both considerations about the inferential process and the common distinctions regarding the types of evidence the inference is based on. Keywords: Inferential evidentiality, evidential inference, evidence, abduction, Jean-​Pierre Desclés, Zlatka Guentchéva, plausibility

18

Dendale and Miecznikowski

1. Introduction The notion of abduction, developed more than a century ago by the logician, semiotician and philosopher Charles Sanders Peirce (1839–​1914), is essential in epistemology and in the formation of new knowledge. However, this notion has not received, we think, enough attention from philosophers of science. (Desclés 2000: 87)

In classifications of markers of evidentiality, inference markers are one of three commonly acknowledged subcategories, beside direct perception markers and reportive/​hearsay markers. They provide evidence of the type “Inferring”, according to Willett’s (1988) classification, and belong to the category named “Inference” by Plungian (2001) and “Inferred” by Aikhenvald (2004). Inferential evidentiality is no doubt the category with the highest number and richest variety of lexical and grammatical expressions. A provisional list of evidential markers (grammatical and lexical) that we made for French in preparation of Dendale (2022) includes at least 33 (acknowledged or suspected) mark­ ers of inferential evidentiality, against only 20 markers of reportive evidentiality and less than a handful of markers of direct perception (most of which are inferential markers as well). Despite the size and importance of the category of inferential evidentials and the growing number of empirical studies on that type of markers, surprisingly little has been written in the literature on evidentiality about the nature of the inferential operation underlying inferential evidentials. As Jean-​Pierre Desclés and Zlatka Guentchéva note, [m]‌any authors, in particular in studies of evidentiality, call upon the notion of inference, without, however, specifying the type of inference used. And yet all inference processes are far from being identical. (Desclés & Guentchéva 2018: 1, our emphasis)

“[T]‌here exist several types of inferences”, states Cornillie (2009: 49). Indeed, what we call ‘evidential inference’ is not just any inference. Its properties and purpose are quite different from the properties and purpose of, for instance, the inference used to compute implicit meanings of utterances, such as conversational implicatures in Gricean and

Inferential Evidentiality and Abduction

19

relevance theoretical pragmatics, or to compute coherence relations between utterances when interpreting texts (Kintsch & Van Dijk 1978, Crothers 1978). Evidential inference may be argued to bear some sim­ ilarity with argumentative inference, i.e. the inference warranting a link between a standpoint and an argument that is spelled out textually (for a theoretical discussion, see Miecznikowski 2020), but the rela­ tions between evidential, argumentative and pragmatic inference are still poorly understood.1 Evidential inference is a notion that needs to be characterized explicitly. How can this be achieved? As far as we can see, two main strategies have been followed by linguists studying evidentiality. The first, and evidently the most common one, is a strategy that focuses on the nature of the evidence on which the inferential operation is based (i.e. on the nature of contextually salient premises), in order to indirectly characterize that operation as a whole. Authors distinguish between perceptual evidence and non-​perceptual (viz. conceptual) evidence (e.g. Diewald & Smirnova 2010: 63, Ruskan 2012, Marín-​Arrese 2013), between perception-​based, conception-​based and communication-​based (or report-​based) inferential readings (Marín-​ Arrese 2017) or between observation-​ motivated knowledge and reasoning-​motivated knowledge (Plungian 2001: 352). This distinc­ tion has given rise to well-​k nown binary oppositions between inferential categories such as “Inferred” vs. “Assumed” (Aikhenvald 2004), “Inferentials” vs. “Presumptives” (Plungian 2001: 353), “(Inference from) Results” vs. “(Inference from) Reasoning” (Willett 1988: 57, 63) or “circumstantial (or specific) inference” vs. “generic inference” (Squartini 2008: 923–​925).2 As we plan to show in subsequent work, this type of dichotomic characterization raises different theoretical problems. What is equally important for the present purpose is that it

1 2

On the relation between pragmatic inference and argumentation, see Rocci (2006) and Oswald et al. (2020). Let us mention also Squartini’s category of ‘pure conjecture’, which the author adds to ‘circumstantial inference’ and ‘generic inference’ (Squartini 2008: 925). Pure conjectures are said to lack “any evidence, both external and based on general world knowledge” (2008: 924). We do not believe, however, that inference is possible without any evidence, viz. without any premises, either major or minor.

20

Dendale and Miecznikowski

fails to provide a general account of what evidential inference precisely is, compared to other types of inference such as argumentation or the computation of implicit meaning. Rooted as it is in typological research about evidentiality as a grammatical category, this approach tends, in fact, to take the specificity of evidential inference for granted on linguistic grounds. Within this approach, what specifically characterizes evidential inference (including all subtypes) is the fact that it is encoded by grammatical evidential markers in a relevant number of languages. Even if a growing number of studies about lexical evidentials has given a new twist to the debate, many linguists still view evidential inference as mainly defined by its opposition to other evidential categories (especially report and perception) within more or less closed paradigms of markers. That may explain the lack, in much linguistic research, of conceptual defining criteria for evidential inference that are independent of its linguistic marking. The second strategy focuses on the nature of the inferential operation itself. The present paper aims at presenting and assessing that strategy, which has been followed by rather few scholars in the field of evidentiality. Jean-​Pierre Desclés and Zlatka Guentchéva (henceforth D&G) are the first scholars, to our knowledge, who have proposed a clear (and radical) definition of the nature of inferential evidentiality –​ inferential mediativity3 in their terminology –​through a characteriza­ tion of the inferential operation itself. The claim they formulated in the mid-​1980s and never stopped defending since is that inferential mediativity is linked, by its nature, to the Peircean notion of abduction,4

3

4

Desclés & Guentchéva, in their articles written in French, use the noun médi­ ativité instead of the more widely used évidentialité. The latter (see Vet 1988, Dendale 1991 for early occurrences) is a calque of English evidentiality and has repeatedly been criticized by D&G and others as being conceptually and linguistically inadequate. Médiativité was transposed into English as mediativity (Lazard 1999, Desclés 2018, Desclés & Guentchéva 2018), indirectivity in Johan­­­ son & Utas (2000) and Johanson (2018). We will discuss the difference between the terms mediativity and evidentiality in section 2. The paternity of the idea to link inferential mediativity to that notion was explic­ itly attributed to Desclés by Guentchéva (1994: 18 fn. 21): “La notion d’abduction m’a été proposée par J.- P. Desclés pour traiter certaines valeurs du parfait bulgare, notamment celle d’inférence et de reconstruction (Guentchéva, 1985; 1990: 181)”.

Inferential Evidentiality and Abduction

21

a type of inference different from both induction and deduction (see section 3). Their focus on the characterization of the entire inferential operation –​or “inferential frame”,5 as we like to call it –​rather than solely on the characterization of single components of it, combined with systematic reference to Peirce’s views on the properties of the different types of inference, is unique in the field of studies about evidential inferentiality. That justifies an attentive reading and synthesis of D&G’s work on abduction, which we will undertake here. We will first show that there are two versions of Desclés & Guentchéva’s claim: a “narrow version” and a “broad version”. In the narrow version, only markers of inferential evidentiality/​mediativity are linked to the Peircean notion of abduction. In the broad version, all types of mediativity/​evidentiality are considered abductive. Abduction then becomes definitional of mediativity as such. We will argue that the narrow version of D&G’s claim raises interesting questions about the nature of mediative/​evidential inference and captures important properties of inferential evidential markers. We are less convinced, however, by the descriptive and explanatory adequacy of their broad claim for the study of markers of evidentiality. After some basic terminological clarifications (section 2) and an introduction to the notion of abduction (section 3), we will present and discuss the narrow claim on the basis of a close reading of more than 40 publications by the two authors (section 4).6 We will, more succinctly, present the broad claim in section 5. Section 6 gives a structured overview of the properties of abduction as

5 6

(‘The notion of abduction was proposed to me by J.- P. Desclés to analyze certain uses of the Bulgarian perfect, especially that of inference and reconstruction’). See also Guentchéva (1994: 9). The frame comprises the input-​evidence, the inferential operation itself, and the output, viz. the resulting thought. We found the term abduction or abductive inference mentioned in at least 44 studies signed by Desclés, Guentchéva or by both authors between 1989 and 2018 (but possibly we did not gather all the relevant studies). The first occur­ rences of the term in their work are to be found in Desclés (1989: 36, 49, note 1) and Guentchéva (1990: 48, note 3). In many of their studies, the term is simply mentioned, sometimes accompanied by a definition and a reference to Peirce. Studies that focus entirely on the notion are: Desclés 1996, 2000, Desclés & Guentchéva, 2001, 2013, 2018, Desclés & Jackiewicz 2006.

22

Dendale and Miecznikowski

they can be distilled from D&G’s writings. A quick overview of issues and unsolved questions raised by D&G’s treatment of the notion will be given in section 7. We hope that this discussion will pave the way toward a fuller understanding of the nature of evidential inference.

2. Evidentiality versus Mediativity: Notional Differences In order to be able to situate D&G’s claim about the nature of (inferential) mediativity/​evidentiality, let us compare the notion of evidentiality with that of mediativity,7 referred to, alternatively, by terms such as médiatif (Lazard 1956: 148, 1996; Guentchéva 1994, 1996: 12), énonciation médi­­­­ atisée (Guentchéva 1990, 1996, Guentchéva & Landaburu 2007), medi­­ ation (Guentchéva 2004), médiativité (Guentchéva 2011, 2013, 2014a), mediativity (Desclés 2018), mediative enunciation (Desclés 2018: 383). In our view, there are at least three important differences between the notions of mediativity and evidentiality. The first difference concerns their subcategories. The typological and discourse analytical literature currently recognizes three main subcategories of evidentiality, termed Attested, Reported and Inferring by Willett (1988) (see § 1). The notion of mediativity, on the other hand, includes the following three subcategories, as exposed in Guentchéva (1996: 16): The category ‘médiatif’ has three fundamental functions: 1. reporting facts that one has learned from a third party or by rumors and hearsay; 2. inferring facts; 3. expressing surprise at the finding of a fact. (our translation)8, 9

7 8

9

Also see Guentchéva’s study (2014a) on the differences between the notions. In the body of the text, all excerpts from D&G’s works are cited in English. When such excerpts in English are our translations from an original French text, the latter is cited in footnotes between double quotation marks. “La catégorie du médiatif s’articule ainsi autour de trois valeurs fondamen­ tales: 1. rapporter des faits dont on a eu connaissance par un tiers, des rumeurs et des ouï-​dire; 2. Inférer des faits; 3. Exprimer sa surprise devant la constatation d’un fait”. Guentchéva (1996: 16).

Inferential Evidentiality and Abduction

23

There seems to be a reasonable overlap between the subcategories Reported and Inferring on one hand and the first two subcategories mentioned by Guentchéva (1996) on the other. The evidential subcategory Attested, however, does not correspond one-​to-​one to the mediative subcategory ‘perception with surprise’. The former implies direct sensorial perception, whereas ‘perception with surprise’ rather seems to amount to ‘mirativity’ (DeLancey 1997, 2001, 2012; Lazard 2001: 361), a notion related to, but different from evidentiality according to many authors (Dendale & Tasmowski 2001: 343, Aikhenvald 2004: 8). A second, essential but less obvious, difference between the two notions concerns the defining features of mediativity. Mediativity is defined by D&G as a form of enunciation, or utterance, as shown by the phrase ‘mediative enunciation’ (énonciation médiatisée). The notion is to be situated in a French linguistic and pragmatic framework, called théorie de l’énonciation “enunciative theory”.10 The theory origi­ nated in the work of the Prague School and of scholars like Pauliny (1967[1948]), Bally (1932/​1965), Benveniste (1966), Culioli (1990, 1999), Desclés (1976) and others. Arguably, its closest counterpart in the Anglo-​Saxon tradition is speech act theory (Desclés 2018: 386). Within the enunciative framework, mediative enunciation contrasts with forms of enunciation such as “epistemic enunciation” (see also section 4.3) and “enunciation of directly perceived facts” (assertion) (Desclés 2018: 397). As the term suggests, mediative enunciation is, in the first place, mediated: While by using ‘ordinary’ (non-​mediative) forms, the speaker states the facts purely and simply, with mediative forms he states them mediately, through the knowledge he has acquired of them, to which he refers and from which he thereby takes a certain distance. […] The speaker is in a certain way split into two persons, one who speaks and one who has heard or concludes or observes. This operation creates a distance between the speaker and his own speech, while in unmarked utterances he adheres to his utterance under the general laws of speech. It is this abstract distancing that constitutes the specific value of the forms in question and which

10 Overviews of that framework are provided by Desclés 2009 and 2018.

24

Dendale and Miecznikowski justifies the denomination “médiatif”. (Lazard, 2000: 212, our translation and emphasis)11

Mediative enunciation implies the speaker’s disengagement (Guentchéva 2004: 31). Unlike plain assertions, it does not commit the speaker to the truth of what s/​he says (Desclés 2018: 386), because in mediative enunciation the speaker only “puts forward a hypothesis while accepting in advance the possibility that it could be contradicted” (Desclés & Guentchéva 2018: 243): in some languages, grammatical units and constructions can signal a "certain disengagement on the speaker’s part". Indeed, by using these units, the speaker simply states a plausible hypothesis and not an assertion (which would fully commit the speaker), or the probable conclusion of reasoning. (Desclés & Guentchéva 2018: 242, our emphasis).

The notion of mediativity thus primarily refers to a form of enunciation that is mediated via something and to disengagement. It, therefore, differs from the notion of evidentiality,12 which is related to “evidence” and to “source of information”, “source of knowledge”, “mode of access to information” or “way of acquiring information”: […] source of the information cannot be retained as a defining criterion of the grammatical category of the mediative, […] and […] the term médiatif cannot be considered as an equivalent of the French term évidentialité. This is

11 “Alors que, en usant des formes ‘ordinaires’ (non médiatives), le locuteur énonce les faits purement et simplement, avec les formes médiatives il les énonce médiatement, à travers la connaissance qu’il en a prise, à laquelle il fait référence et par rapport à laquelle il prend par là même un certain recul. […] Le locuteur est en quelque sorte scindé en deux personnes, celle qui parle et celle qui a ouï-​ dire ou qui conclut ou qui constate. Cette opération crée une distance entre le locuteur et son propre discours, alors que dans l’énoncé non marqué il adhère à son dire en vertu des lois générales du discours. C’est cette distanciation abstraite qui constitue la valeur propre des formes en question et qui justifie la dénomination de ‘médiatif’ ” (Lazard, 2000: 212, our emphasis). 12 See, however, Palmer, who considered evidentiality as “indication by the speaker of his (lack of) commitment to the truth of the proposition being expressed” (1986: 51).

Inferential Evidentiality and Abduction

25

not a mere terminological problem, but a methodological and epistemological problem. (Guentchéva 2014a: 70, our translation)13

A third difference between the two notions, which we will discuss below (section 4.3), is that mediativity is linked to the plausibility of the communicated content, a notion that is less central in the theoretical discussion about evidentiality.

3. The Peircian Notion of Abduction according to D&G’s Presentation Peirce himself defines abduction, which he also calls retroduction (CP, 1.65),14 hypothesis (1.65 fn) or presumption (2.777, 2.774), in the follow­ ing terms (CP 5.189): The surprising fact, C, is observed. But if A were true, C would be a matter of course, Hence, there is reason to suspect that A is true

In Desclés & Guentchéva’s wording, abduction is a form of reasoning “based on facts (observed or known) and the law of inference (conditional relation or material implication) across propositions that states the plausibility of a hypothesis” (Desclés & Guentchéva 2018: 241–​ 242), or, put more formally: Abductive reasoning is reasoning “based on observed facts ‘F’ which lead to a plausible hypothesis ‘H’, on condition that, in addition, the one entails the other: ‘H =​> F’. Inferential abductive reasoning, ‘H =​> F’ & ‘F’ therefore ‘H is plausible’ is distinct from inferential deductive reasoning: ‘H =​> F’ & ‘H’ therefore ‘F’ ”. (Desclés, 2016: 31, note 1) 13 “la source de l'information ne peut pas être retenue comme critère fondant la catégorie grammaticale du médiatif, […] et […] le terme médiatif ne peut pas être considéré comme un équivalent du terme français évidentialité. Il ne s'agit donc pas d'un simple problème terminologique mais d'un problème métho­ dologique et épistémologique. ” (Guentchéva 2014a: 70). 14 The numbers indicate the volume, followed by the paragraph in the Collected Papers (CP) of Peirce.

26

Dendale and Miecznikowski

Three elements are definitional in abductive inference: observed facts F, a plausible hypothesis H and a general law, which takes the form of an implication “H =​> F” and states a relation between the hypothesis and the observed facts. Another important point concerns the semantic relation between the states of affairs described by H and F. According to the following quotes, this relation is causal: Abduction uses a general law and an observation to find a ‘causal hypothesis’ (Desclés & Jackiewicz 2006: 36, our translation).15 Through abduction, in a certain sense one retrieves a cause starting from an observed fact. (Guentchéva 1990: 49 fn.3, our translation)16

This position seems coherent with the fact that in most examples presented by D&G, the mediative marker qualifies a statement about a possible cause of the observed facts. On the other hand, the authors point out that a broader range of semantic relations can be observed: A conditional "if p then q" will be accepted as a general law when it expresses a relation between a cause and one of its effects. Yet, we must not think that all reasoning by abduction is linked to a relation of causality. Indeed, certain abductive reasonings implement a general rule of praxeological nature17 that no longer expresses a relationship between a cause and an effect or one of its symptoms. (Desclés & Guentchéva 2001: 115, our translation and empha­ sis)18

15 “L’abduction utilise une loi générale et un constat pour remonter à une ‘hypothèse causale’ ”. (Desclés & Jackiewicz 2006: 36). 16 “Par abduction, on remonte d’une certaine façon à une cause à partir d’un fait constaté”. (Guentchéva 1990: 48 fn.3). 17 Praxeological refers to the rule-​governed practice of deriving acceptable con­ clusions from a set of premises: a praxeological rule “allows one to act and to infer” (“permet d’agir et d’inférer”, Desclés & Guentchéva 2001: 108, fn. 10). 18 “Une conditionnelle ‘si p alors q’ sera acceptée comme loi générale lorsqu'elle exprimera une relation entre une cause et un de ses effets. Il ne faudrait pas penser que tout raisonnement par abduction serait lié à une relation de causalité. En effet, certains raisonnements abductifs mettent en œuvre une règle générale, de nature praxéologique qui n'exprime plus une relation entre une cause et un effet ou un de ses symptômes” (Desclés & Guentchéva 2001: 115, our emphasis).

Inferential Evidentiality and Abduction

27

­ e may conclude that for D&G, abduction is typically, but not excluW sively, causal. Abduction is one of the three types of inference distinguished by Peirce, besides the much better-​k nown operations of deduction and induction. The difference between the three types is the following: Induction generalizes, deduction infers a conclusion from a premise assumed to be true, abduction infers the plausibility of a hypothesis from an observation. (Desclés & Jackiewicz 2006: 37, our translation and emphasis)19

In section 6, we will present in more detail the components of abductive inference as it is defined and used by D&G.

4. The Narrow Version of D&G’s Claim 4.1. What is It? The narrow version of D&G’s claim –​which is its earliest version and the one that is presented in most of these authors’ writings,20 –​links inferential mediativity/​evidentiality to the Peircean notion of abduction or abductive inference (cf. Dendale & De Mulder 1996: 305). The oper­ ation of inference referred to by inferential evidentials (viz. evidential inference in our terms) is essentially abductive according to D&G: Inference by abduction is expressed in many languages such as Bulgarian, for example, where the perfect form has led to the creation of a system of perfectoid forms that have specialized in expressing this value. (Guentchéva 1994: 19, our translation and emphasis)21

19 “L’induction généralise, la déduction infère une conclusion à partir d’une prémisse supposée vraie, l’abduction infère la plausibilité d’une hypothèse à partir d’un constat”. (Desclés & Jackiewicz 2006: 37, our emphasis). 20 All examples given in Desclés & Guentchéva 2000 and 2018 are only of the inferential type. 21 “L'inférence par abduction trouve une expression linguistique dans de nom­ breuses langues comme le bulgare, par exemple, où la forme du parfait a conduit

28

Dendale and Miecznikowski

In the narrow version, non-​inferential markers seem to be excluded from being abductive, as shown in the following passage, where “abductive reasoning” is put on a par with a report by an “unspecified third party”, “hearsay” and “unexpected observation”, with the adjective abductive only applied to reasoning (i.e. to inference), not to the other categories. [The notion of mediation] refers to a complex act of enunciation, which consists in presenting facts whose apprehension […] corresponds […] to a "mediated report" the enunciator establishes between him/​herself and the propositional content of his/​her utterance; the latter can be attributed to an unspecified third party, to hearsay, to an abductive reasoning based on clues generally derived from perception, or even to facts involving an unexpected observation. (Guentchéva 2004: 25, our translation and emphasis)22

But even in the narrow version, D&G’s claim is radical: a language item, according to them, is an inferential mediative or evidential marker if and only if it expresses an inference that is abductive. 4.2. Examples of Mediative/​Evidential Markers of Abductive Inference In this section, we illustrate abductive inference with a few examples taken from D&G’s work. According to Guentchéva, the French passé composé (cp. Present perfect) as it is used in sentences (1) and (2) (bold face) can “only be analyzed as having an inferential value” (1994: 19). Indeed, in the specific context of these utterances, this tense-​aspect marker acquires

à la création d'un système de formes perfectoïdes qui se sont spécialisées dans l'expression de cette valeur”. (Guentchéva 1994: 19, our emphasis). 22 “[La notion de médiation] renvoie à un acte complexe d'énonciation qui con­ siste à présenter des faits dont l'appréhension […] correspond […] à un ‘rapport médiat’ que l'énonciateur institue entre lui et le contenu propositionnel de son énonciation, ce dernier pouvant être attribué à un tiers non spécifié, à un ouï-​dire, à un raisonnement abductif à partir d'indices généralement issus de la perception, ou encore à des faits impliquant un constat inattendu”. (Guentchéva 2004: 25, our emphasis).

Inferential Evidentiality and Abduction

29

a mediative/​ evidential value, more specifically an “Inferring” one, according to Willett’s (1988) classification: (1)

Regarde les yeux rouges du concierge ! –​ Il a pleuré ! –​Non, il a dû boire. [‘Look at the janitor’s red eyes! –​ ‘He has cried! –​‘No, he must have been drinking’].

(2)

–​Tiens ! La valise de François n’est plus dans sa chambre! [‘Look at that! Frank’s suitcase is no longer in his room!’] –​ Il est parti !    [‘He is gone’] –​Non, il a dû la déplacer [‘No, he must have moved it’] (Guentchéva 1994: 19)

For D&G, the two sentences in the passé composé express conclusions arrived at by abductive inference. They both start from an observed fact (the janitor’s red eyes; Frank’s suitcase of Frank that is no longer in his room) and explain those facts causally by plausible hypotheses: that the janitor has been crying and that Frank has gone. In both examples, the formulated hypothesis is countered by another speaker’s utterance that contains an even more explicit inferential marker, epistemic devoir (‘must’) and puts forward an alternative hypothesis: that the janitor, plausibly, has been drinking23 and that Frank, plausibly, has moved his suitcase to another place. These alternative hypotheses are the results of abductive inferences, too: they are based on the same observed facts and lead to hypotheses that are also merely plausible (albeit ‘more plausible’ for the replying speaker than the first set of hypotheses). Throughout their work, Desclés & Guentchéva illustrate the marking of abductive inference with examples taken from French,24 23 Cf. ‘each interlocutor expresses his/her opinion on the plausible cause of the redness of the janitor’s eyes’ “chaque interlocuteur exprime son avis sur la cause plausible de la rougeur des yeux du concierge” (Guentchéva 1994: 19). 24 In their work, we found the following French markers being associated in one way or another with abductive inference: the present perfect, the inferential

30

Dendale and Miecznikowski

­Bulgarian25 and a series of other languages, drawing on the broader typological literature. The following example, commented by Guentchéva (2004), concerns the inferential marker -​re .in Wintu: (22) ničaj ɂewin suke-​ re . nephew here stand-​I NFERENTIAL [‘My nephew must have been here (see tracks)’] By -​re ., in (22), the enunciator signals that he is inferring a plausible past fact from present clues. This can be associated with the Peircean operation of abduction (II: 636; V: 171). (Guentchéva 2004: 26, our translation)26

4.3. Markers of Abduction and Markers of Deduction The inferential expressions exemplified in the previous section are markers of abductive inference. Since 2001, however, Desclés & Guentchéva (2001), in reply to criticisms advanced by Dendale & De Mulder (1996), defend the idea that inferential expressions, in specific configurations, can also mark a deductive inference. One might ask how this idea is compatible conditional in yes-​no questions (Guentchéva 2003: 180), epistemic devoir, epis­ temic pouvoir (Desclés 2003: 53), the adverbs sans doute (Desclés 2009: 43) and certainement (Desclés & Jackiewicz 2006: 39–​40). Even the conclusion marker donc, combined either with a passé composé, the conditional mood (Guentchéva 2003: 180; Desclés & Guentchéva 2018: 250–​251) or with epistemic devoir, is considered a marker of abductive inference in several publications by the authors (e.g. Guentchéva 2003: 180; 28; Desclés & Jackiewicz 2006: 37; Desclés, 2009: 42–​43). We agree with the mainly abductive nature of all the cited mark­ ers, except for donc, which for us is a conclusion marker that can perfectly mark deductive inferences (but we cannot develop further our argumentation on that point). For some remarks about the difference between inferential connectors and evidential inferential markers, see Miecznikowski (2015: 104). 25 E.g.: ‘The notion of abduction is of great importance for the understanding of a large number of inferential uses of the Bulgarian perfect.’ “La notion d'abduction est d'une grande importance pour comprendre un grand nombre d’emplois du parfait bulgare à valeur inférentielle”. (Guentchéva 1990: 48, n. 3). 26 “Avec -​re ., en (22), l'énonciateur signale qu'il infère un fait antérieur plausi­ ble à partir d'indices présents. On peut l'associer avec l'opération d'abduction de Peirce (II: 636; V: 171)”. (Guentchéva 2004: 26).

Inferential Evidentiality and Abduction

31

with the narrow version of their claim stating that all inferential markers are abductive in nature. At the end of this section, we will see how D&G reconcile the idea with that claim. That abductive markers should be distinguished from deductive markers is argued e.g. by D&G (2013): It is important not to confuse the linguistic expression of deductive reasoning (by modus ponens) leading to a probable consequence with abductive reasoning leading to a plausible hypothesis. […] We will explicitly distinguish abductive reasoning from deductive reasoning with a probabilized consequence, which leads to probable conclusions (Desclés & Guentchéva 2013: 81–​82, our translation and emphasis).27

To refer to deductions that do not lead to a necessary conclusion, different wordings are used in D&G (2001): “déduction avec un conséquent jugé probable” (2001: 104), “raisonnement déductif avec un conséquent probabilisé” (2001: 115); in D&G (2018: 250) the operation is called “deduction with a probable consequent”. The concrete interpretation of a marker –​as either abductive or deductive with a probable consequent –​depends on the content and the context of the utterance. For D&G, the need to distinguish between those two types of markers arose because of examples like (4) advanced by Dendale & De Mulder (1996: 305), as opposed to “abductive” exam­ ples like (3). These examples were used by Dendale & De Mulder to call into question the claim that evidential inferences are necessarily abductive (as proposed in Guentchéva 1994), leading from effect to a cause, and to defend instead –​at the time –​a deductive analysis of such inferences, leading from a cause to its effect:

27 “Il est important de ne pas confondre l'expression linguistique des raisonne­ ments déductifs (par modus ponens) aboutissant à une conséquence probable avec les raisonnements abductifs aboutissant à une hypothèse plausible. […] Nous distinguerons explicitement le raisonnement abductif du raisonnement déductif avec conséquence probabilisée, conduisant alors à des conclusions probables”. (Desclés & Guentchéva 2013: 81–​82, our emphasis).

32

Dendale and Miecznikowski

(3)

Jean est fatigué. Il doit avoir travaillé beaucoup (Il a dû travailler beaucoup)28 aujourd’hui. [‘John is tired. He must have worked a lot today’]

(4)

Jean a beaucoup travaillé aujourd’hui. Il doit être fatigué. [‘John worked a lot today. He must be tired’].

Both examples contain a form of the epistemic verb devoir “must”, which Dendale & De Mulder (1996) consider an inferential evidential marker in both cases. The semantic relations between the successive sentences, however, are different in (3) and in (4). In (3), the sentence containing devoir indicates the causal explanation of the psycho-​physical state (He is tired) described in the first sentence. In (4), on the contrary, the sentence with devoir indicates the consequence of the event described in the first sentence. In both sequences, there is a causal relation between the sentences, but the direction of causality is opposite.29 For Desclés & Guentchéva (2001), devoir in (3) is an abductive marker. From an observed fact (John is tired), the speaker retrieves a causal explanation (John has worked a lot today), warranted by a known “general law” (If someone works a lot, s/​he is probably tired (2001: 116)). The explanation is presented as a plausible (but uncertain) hypothesis concerning the observed fact. In contrast, the abductive analysis evidently does not hold for (4), although there is an observable fact, too (John worked a lot today), and the clause with devoir expresses an uncertain hypothesis, just like in (3). Since devoir does not mark 28 In French, both constructions are equally possible, with roughly the same mean­ ing. In English, there is no equivalent of the second one, i.e. of the epistemic verb being used in the present perfect in combination with an infinitive. 29 The possibility to argumentatively exploit semantic relations in two directions was theorized systematically in medieval treatises such as Peter of Spain’s Summulae logicales, which heavily influenced the subsequent tradition of studies in rhetoric and argumentation (cf. Rigotti & Greco 2019: 93–​204). The theory is still echoed, in contemporary practice and teaching, by common symmetric pairs of denominations for argument schemes and rhetorical figures such as the causal pair “from cause to effect” and “from effect to cause”, which we have applied to the examples (3) and (4), or the mereological pair “from the parts to the whole” and “from the whole to its parts”.

Inferential Evidentiality and Abduction

33

an inferred cause here, but an inferred consequence, it functions as a marker of prediction rather than as a marker of causal explanation. Given D&G’s strong inclination to identify abduction with the finding of a causal explanation, they do not consider devoir in (4) the sign of an abductive inference. They, therefore, concede that devoir in (4) is the sign of deductive inference, adding –​as anticipated at the beginning of this section –​the specification that this kind of deduction has “a probable consequent” (Desclés & Guentchéva 2018: 250). This means that the “deductive reasoning” (2018: 248) underlying devoir in (4) rests on a “law of inference”, a “conditional relation” or “material implication” whose consequent is only probable (2018: 241), (see (5)): (5) p =​> prob (q) (Desclés & Guentchéva 2001: 116)

The probabilized law or conditional relation underlying (4) is claimed to be something like (6): (6) Si quelqu’un travaille beaucoup, alors il est probablement fatigué. (Desclés & Guentchéva 2001: 116) ‘If someone works a lot, then s/​he is probably tired’.30

One obtains the following deductive reasoning with probable c­ onclusion: (7)

1. Jean a beaucoup travaillé. (premise) 2. (∀x) [travailler-​beaucoup (x) =​> prob (être-​fatigué (x))] (law) 2’. travailler-​beaucoup (Jean) =​> prob (être-​fatigué (Jean)) (instantiation) -​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​-​ 3. prob (être-​fatigué (Jean)) 3’. Enunciation: Jean doit être fatigué. (Desclés & Guentchéva 2001: 116)

The difference between abductive inference and deductive inference can be presented schematically as follows (after Provôt & Desclés 2012: 48, our translation):

30 One could also conceive that what is at stake here is “default reasoning” (Reiter 1978) of the type: “If someone works a lot, s/​he generally/​normally is tired” or “People who work a lot, are generally/​normally tired”.

34

Dendale and Miecznikowski

Table 1.  Abductive and deductive schemes of reasoning Type of inference:

Deductive

Abductive

Major Premise: Rule, Law Minor Premise: Fact ------------------------------Conclusion: Case

if p then (possibly / probably) q p -----------------------------------(possibly / probably) q

if p then q q ---------------------p is plausible but p’ can contradict p

This solution, proposed by D&G for the problem addressed by examples (3) and (4), has an important consequence for the description of the meaning of devoir, and of inferential markers in general. Maintaining that all inferential mediative (viz. evidential) markers signal abductive inference, D&G consider epistemic devoir an evidential (mediative) marker when used as in (3) but not when used as in (4). In (4), they analyze it as an epistemic-​modal marker, in which case it contributes to what they call “epistemic enunciation” (Desclés 2018: 397; Desclés & Guentchéva 2013: 95), as opposed to “mediative enunciation”. The term epistemic enunciation, for the authors, can refer to two things: The speaker can also resort to an epistemic enunciation. The latter then indicates that his/​her judgment is situated in the sphere of uncertainty (She must have left, it is a possibility) and the speaker does not seek to indicate that there are reasons (the observation of clues, for example) justifying the refusal of his/​ her full commitment to the predicated content. In some contexts, epistemic utterances also indicate a probable consequence established as a result of deductive reasoning, based on generalizing implications. (Desclés & Guentchéva 2013: 95, our translation)31

31 “L’énonciateur peut également avoir recours à une énonciation épistémique. Cette dernière indique alors que le jugement se situe dans l’incertain (Elle a dû partir, c’est une possibilité) et l’énonciateur ne cherche pas à indiquer qu’il y aurait des raisons (reconnaissance d’indices, par exemple) susceptibles de justifier le refus de son plein engagement dans la prise en charge d’un contenu prédicatif. L’énonciation épistémique indique également, dans certains contextes, une conséquence probable établie, cette fois, à la suite d’un raisonnement déductif, fondé sur des implications généralisantes”. (Desclés & Guentchéva 2013: 95).

35

Inferential Evidentiality and Abduction

Epistemic enunciation can amount either to an epistemic modal judgment of certainty, without any indication of a source of knowledge, or to a deductive inference with a probable consequent and thus with a probable conclusion. Devoir can be used in both types of epistemic enunciation. The context –​more specifically, the type of relation between the clues and the conclusion –​is said to disambiguate between the two interpretations of “epistemic” devoir (Desclés & Guentchéva 2001: 121), and to determine the status of the marker, according to the parameters summed up in Table 2. Table 2.  Properties of epistemic devoir in examples (3) and (4) Epistemic devoir Example (3)

Example (4)

Abductive inference

Deductive inference

Plausible conclusion

Probable conclusion

Evidential/​mediative marker

Epistemic modal marker

The introduction of “deduction with a probable consequent” as an epistemic modal category has one main objective: to save the claim about the generalized abductive nature of mediative inference by excluding expressions of deductive inference from the category of mediative (evidential) markers and qualifying them as epistemic markers instead. 4.4. The Double “Domain of Use” of the Notion of Abduction in D&G’s Work In the examples given in sections 4.2 and 4.3, we have presented the way abduction, or the “abductive inference process”, is referred to by linguistic expressions (see the title of Desclés & Guentchéva 2018) –​ both grammatical (Desclés & Guentchéva 2013: 81) and lexical (Provôt & Desclés 2012: 47). What has thus been in focus, up to now, is

36

Dendale and Miecznikowski

how abduction relates to the linguistic phenomenon of evidentiality/​ ­mediativity. Abduction is, of course, more than just a linguistically marked phenomenon related to mediativity/​ ­ evidentiality. It corresponds, more generally, to a form of inference that is “essential in scientific ­reasoning”,32 (Desclés 2003: 53, fn. 10), “in epistemology and in the for­­ mation of new knowledge” (Desclés 2000: 87). Abductive reasoning is present in medical diagnostics, historical, archeological and geological reconstruction, detective stories and ordinary language conversation: Yet, abduction is employed in spontaneous reasoning and also in a large number of scientific reasoning (including mathematics). Man in this way has made it possible for himself to state hypotheses that refer to some not directly observable entities. (Desclés 2000: 87). This is why judges, police officers, physicians, archaeologists, geologists, evolutionists, historians, comparative linguists who reconstruct a –​not directly observable –​state of language all have recourse to abductive reasoning in their practices and will seek to accumulate relevant clues, by carrying out additional analyses, by triggering new observations, by researching new documents, in order to better support a hypothesis they have put forward and which will thus gradually become an "explanatory cause" of the phenomena observed and described. (Desclés & Jackiewicz 2006: 39, our translation).33

32 Including linguistics as a scientific discipline. To illustrate the role of abduction in linguistics, Desclés & Guentchéva (2013: 218) mention Ferdinand de Sauss­ ure’s reconstruction of proto-​Indo-​European phonemes on the basis of known phonetic laws and the decipherment by Champollion of Egyptian hieroglyphs based on the Rosetta stone translations. 33 “C’est pourquoi les juges, les policiers, les médecins, les archéologues, les géo­ logues, les évolutionnistes, les historiens, les linguistes comparatistes qui reconstruisent un état de langue –​non observable directement –​, qui ont tous recours, dans leur pratique, à des raisonnements abductifs, chercheront à accumuler les indices pertinents, en procédant à des analyses complémentaires, en déclenchant de nouvelles observations, en recherchant de nouveaux documents, pour mieux étayer une hypothèse avancée qui deviendra ainsi, progressivement, une ‘cause explicative’ des phénomènes constatés et décrits” (Desclés & Jackiewicz 2006: 39).

Inferential Evidentiality and Abduction

37

Throughout their publications, D&G use the term abduction to refer alternately to the operation of which mediative/​evidential markers are “traces” (as they call it), or to the reasoning they practice themselves as linguists to achieve one of the following aims: the explicit description of the grammatical meaning of observable forms (Desclés 1989: 36), the interpretation of linguistic units and of linguistic discourse (like the Rosetta stone, Desclés & Guentchéva 2013: 218), the formulation of an invariant meaning for a language item, the formulation of universals of language (Desclés & Guentchéva 2012: 124): […] mental representations and mental operations are constructed by abduction from identifiable clues (traces). (Desclés 1995: 10, our translation and emphasis)34 While universals are language categorizations that transcend specific languages, invariants are abstractions based on attested variation. Therefore they [universals and invariants] are not simple inductive generalizations, but instead are the result of abductive processes in the sense of Peirce (1965, 1967). (Desclés & Guentchéva 2012: 124, our emphasis)

Other applications of abduction in linguistics mentioned by D&G are the formulation of hypotheses about the mental operations behind a linguistic unit (Desclés 1995: 10), the formulation of hypotheses on the evolution of a linguistic item,35 the conception of semantic maps and “network[s]‌of abstract concepts” (Desclés 2016: 30–​31), etc. The very first mention of the term abduction we found in D&G’s work (more specifically in Desclés 1989: 36, 49, note 1),36 actually concerns the

34 “les représentations mentales et les opérations mentales sont construites par abduction à partir d’indices (traces) identifiables”. (Desclés 1995: 10, our emphasis). 35 In the same vein, Deutscher (2002: 469) observes: “The concept of abduction […] was introduced into linguistics almost thirty years ago by Henning Ander­ sen (1973), and has become extremely popular and widespread in the last three decades, especially in historical linguistics. Andersen was not the first linguist to mention the term abduction. Chomsky (1968 [1972]: 90 ff.), for example, discussed abduction in the context of his claims about innateness”. (Deutscher 2002: 469). 36 Indirect references in Guentchéva (1994) further trace back until at least 1985 (cf. our footnote 4).

38

Dendale and Miecznikowski

abductive activity of the linguist rather than the notion behind a mediative marker: In fact, a linguist must follow explicit abductive acts and start from observable forms in order to construct grammatical meanings and represent them. (Desclés 1989: 36, our translation and emphasis)37

It is, of course, D&G’s thinking about the abduction in relation to the phenomenon of mediativity/​evidentiality that is primarily of interest to us here. Their more general “epistemological” understanding of the notion is relevant here, among others, because it might have played a role in the development of what we call the “broad version” of D&G’s claim on evidential inference, to which we turn in the next section.

5  The Broad Version of D&G’s Claim 5.1. What is It? The broad version of D&G’s claim amounts to saying that all subcategories of mediativity, all types of markers of mediativity, and the underlying mediative enunciation as such, are fundamentally abductive in nature: “abduction is central in the analysis of mediated enunciation” (Desclés 2000: 53, our translation).38 This means that not only inferential markers of mediativity/​evidentiality are abductive (as said in the narrow version of their claim) but also markers of reportive mediativity (taken as a general term covering hearsay, quotative…) and even markers of what the authors call “perception with surprise”-​mediativity. This extension is consistent

37 “En fait, un linguiste doit, à la suite de gestes abductifs explicites, partir des formes observables pour construire les significations grammaticales et les représenter”. (Desclés 1989: 36, our emphasis). 38 “l'abduction est centrale dans l'analyse de l'énonciation médiatisée (Guentchéva 1990, 1994, 1996)” (Desclés 2000: 53).

Inferential Evidentiality and Abduction

39

with D&G’s view of abduction as a fundamental and pervasive form of everyday (and scientific) reasoning. 5.2. The Broad Version and the Analysis of Reportive Markers Let us illustrate the broad version of D&G’s claim by comparing two examples, one containing the inference marker devoir “must” and the other containing the French conditional in its reportive use: (8) (9)

L’assassin de Mme Hobson doit être M. Smith. ‘The murderer of Mrs. Hobson must be Mr. Smith.’ L’assassin de Mme Hobson serait M. Smith. ‘The murderer of Mrs. Hobson is said to be Mr. Smith.’

We agree with D&G that devoir in (8) is an inferential evi­ dential. Its function is to indicate that the sentence containing that verb expresses a plausible content inferred on the basis of some other information (for example, Mrs. Hobson’s blood found on the clothes of M. Smith) and a “general law”. If one accepts D&G’s narrow claim that the inference that underlies inferential evidentials is abductive, the reasoning in (8) can be represented as follows, applying the schema presented in the third column of table 1 (sec­ tion 4.3.): (10) q Mrs. Hobson’s blood was found on the clothes of Mr. Smith. if p then q If A kills B, the blood of victim B is (often) found on the clothes of A. Thus, p is Thus, Mr. Smith (A), plausibly, killed Mrs. Hobson (B). plausible (or plausibly p)

The conditional form in (9), in contrast, is not an inferential, but a reportive marker, according to existing analyses of that reading of the French conditional form (cf. inter alia Dendale 1993, 2018). Its function is to indicate that the information in the sentence (Mrs. Hobson’s killer

40

Dendale and Miecznikowski

is Mr. Smith) was obtained by the speaker through other people’s sayings. Provôt & Desclés (2012), however, identify in this use elements of abductive inference comparable to those postulated for inferential devoir: 39 The ‘conditional of rumor’: This type of conditional is mediative because it refers to “sayings” that constitute the clues of an abductive reasoning, which takes the form: “everyone says p40; if a fact p occurs, then every­ one will talk about it, so p is plausible”. (Provôt & Desclés 2012: 49, our translation)41

The abductive reasoning that Provôt & Desclés claim to be at work here goes as follows:42

39 Their analysis is consistent with the way Desclés & Guentchéva treat the French conditional form also in other publications. D&G’s examples of this reading of the conditional form, described in inferential terms, are often in contradiction with its rather generally acknowledged analysis as a reportive marker. Consider e.g. (D’après ce que je sais ou ai pu reconstituer) le président aurait reçu des diamants (Desclés & Guentchéva 2000: 99) “(From what I know or was able to reconstruct) the president [has]CONDITIONAL received diamonds”. In this constructed example, the context specifies clues that are typical of an abductive inference rather than of a report. In many cases, the authors add the conclusion marker donc “thus” to the conditional and thereby orient its interpretation toward inference rather than report: Donc, la mère serait coupable “Thus, the mother [is]CONDITIONAL guilty.” (Desclés 2016: 80, our emphasis). 40 It is significant that the premise contains the universal quantifier “everybody” and not a “someone”, which is also a possibility for reportive markers, but which would not lead to the same plausibility. 41 “Le conditionnel de ‘rumeur’: Ce type de conditionnel est médiatif, car ce sont des ‘dires’ qui constituent ici les indices d’un raisonnement abductif, celui-​ci se présentant alors sous cette forme: ‘tout le monde dit p, or si un fait p se produit, alors tout le monde va en parler, donc p est plausible’ ”. (Provôt & Desclés 2012: 49). 42 For a comparable analysis, see also Desclés (2016: 80). We did not find a similar analysis in work published by Guentchéva (or co-​authored with her). See e.g. the more nuanced analysis of the French conditional in Guentchéva. (2014b: 43).

Inferential Evidentiality and Abduction

41

(11) q if p then q Thus, p is plausible (or: plausibly p)

Everybody says that Mr. Smith killed Mrs. Hobson. If an event E takes place, then everybody will talk about it.43 Thus, the event of Mr. Smith killing Mrs. Hobson (E), plausibly, occurred.

Thus presented, the analyses (10) and (11), which describe the examples (8) and (9), respectively, appear to have comparable abductive inference schemes. In both cases, there is an “observed fact”, a “general law” (which is conditional and causal), and a conclusion, estimated to be plausible. The apparent parallelism between the abductive schemes (10) and (11), however, conceals important differences between the two cases and, in our view, leads to undesirable conclusions about the nature of inferential and reportive markers. A first difference concerns the semantic complexity of q (the “observed fact”, which also appears in the consequent of the “conditional law” if p then q). In the abductive schema proposed for reportive markers, (11), q comprises an explicit element of report/​r umor (Everybody says that…) with scope over an embedded proposition describing a fact E. In the abductive schema (10), in contrast, q directly regards a fact E. A second difference, tightly related to the first, concerns the relationship between the abduced conclusion p and the “observed fact” q. Propositions p and q describe distinct events in (10), while in (11) proposition q contains p. This situation is characteristic of all reports: the speaker integrates in his/​her own discourse an idea that someone else has brought up before. Linked to that, one might argue that the information that is inferred in the conclusion of (11) is not p but rather a judgment of plausibility regarding p. In the two cases illustrated by (10) and (11), not only p but also the plausibility judgment about p thus have a different status. A third difference concerns the content of the major premise or “conditional relation” in the two schemas. In (10), the relation between

43 Or shouldn’t this rather be a probabilized law, with probable in the conse­ quent: everybody will probably talk about it?

42

Dendale and Miecznikowski

antecedent and consequent results from material causal laws, whereas in (11) it corresponds to a “social regularity” (“people tend to talk about things that happen”). The law in (11) implies a chain of human experiences and actions (people experiencing E, people starting to talk about E, people learning about E thanks to others’ discourse and spreading the word), which is subject to numerous intervening factors: for instance, people not having access to E, people distorting the information while passing it on –​because of a lack of expertise, or for other reasons –​, people deciding not to talk about E, etc. Such factors could interrupt the spread of the information or result in a proposition that does not correspond to E, i.e. that is not true. The literature on arguments from authority (e.g. Hastings 1962: 135–​138, Walton 1997, Rigotti 2009, Rocci 2017: 51–​59, 395–​396, Rigotti & Greco 2019: 252–​ 253, 264) rightly suggests that the conceptual link between an event and the discourses produced about it is more complex than the link between causes and effects. The parallelism constructed by D&G between (10) and (11) ignores the peculiarities of the two schemes. It results in an analysis of reportive markers that overly downplay their difference from inferential markers and blurs the important distinction between the features “Other” and “Self” (Frawley 1992: 412), widely accepted in typological and descriptive studies on evidentiality. It de facto reduces the category “Reported” to the category “Inferring” and fails to capture the main function and characteristic of reportive markers, which is to signal that the content of a given utterance comes from somebody else (“Other”) and –​contrarily to inferences –​does not originate in the speaker’s own thoughts (“Self”). The conflation of report with (abductive) inference also makes it impossible, by the way, to grasp the subtle distinction between what is to be analyzed as reportive evidential marking in (9) and what is to be analyzed as inferential evidential marking on the basis of verbal clues in (12): (12) Me basant sur les rumeurs qui circulent, l’assassin de Mme Hobson doit être M. Smith. [‘Basing myself on rumors that are circulating, the murderer of Mrs. Hobson must have been Mr. Smith.’]

Inferential Evidentiality and Abduction

43

One aspect that D&G’s analysis of reportive markers is meant to explain is that the reported content, just like content marked as inferred, represents information that is plausible but does not commit the speaker to its truth: Other clues may be involved in grammaticalized abductive inference, e.g. hearsay (‘a large number of people agreeing on something is a clue in favor of the plausibility of what they agree on’), the words of an indeterminate third party who may have been the direct witness to an event which thereby becomes plausible and is presented as such by the enunciator, despite not having been a witness to the event in question. The enunciation of a plausibility sanctioned by intermediate witnesses bears the technical label "mediative enunciation" and contrasts with the direct statement of sensory or cognitive observations which directly commit the enunciator (‘I’m saying this because I saw/​heard/​observed it for myself …’). (Desclés 2018: 395) In having recourse to mediative enunciation, the enunciator avoids taking full responsibility for the contents of what is being presented as plausible, thus allowing for the possibility of other (or better) hypotheses in lieu of the one expressed, all while explaining the observed clues. (Desclés 2018: 395)

It is true that both reportive and inferential markers tend to weaken the speaker’s epistemic commitment to the utterance content (on the question of commitment in the case of the reportive conditional in French, see Dendale 2018 and Desclés & Guentchéva 2000: 96). However, we tend to think that the origin of weak commitment differs in the case of markers such as the reportive conditional and in the case of inferential markers like devoir. The former lead to weak commitment mainly because p originates outside the speaker, thereby creating an effect of distancing.44 Inferential markers like devoir, on the other hand, reduce commitment because the inferential process is presented as not completely reliable, nonmonotonic, and therefore defeasible (see Dendale 2020: 47, 53). By making mediativity converge with plausibility, D&G’s broad claim fails to account for these differences and further drives away the notion of mediativity from that of evidentiality, whose conceptual core is source of information. 44 On the relation between the speaker’s degree of closeness to what is said and the reduction of enunciative responsibility, see also Caffi’s (2007) analysis of mitigating devices.

44

Dendale and Miecznikowski

5.3. The Broad Version and the Analysis of “Unexpected Finding” Markers An abductive analysis is also proposed by D&G for what they call markers of “state of surprise” or “unexpected finding”:45 “L’état de sur­ prise s’explique également à partir d’un raisonnement abductif” (‘The state of surprise, too, can be explained by an abductive reasoning’), (Desclés & Guentchéva 2003: 53). In an example given by Desclés & Guentchéva (2000), the starting point of the abductive reasoning is the finding of a fact that contradicts what the speaker expected or previously knew: Let’s take the following example: “Look at that, it has rained”. Here, the abductive reasoning is the following: 1°) The speaker knows that if it rains, the roadway is wet, and his/​her state of knowledge (or belief) is: “the weather today was fine”; 2°) the speaker observes that the road is wet; 3°) by abduction, he/​she formulates the hypothesis ("it has rained") and must then revise his/ her state of knowledge; this motivates his/her utterance, which signals his/her change of knowledge. The result is an effect of astonishment, admiration… (Desclés & Guentchéva 2000: 101, our translation)46

The analysis proposed here for the marking of unexpected findings is very similar to the analyses D&G propose for utterances containing inferential abductive markers (see section 4.2.). This is not surprising, because their example (Tiens, il a plu) contains a present perfect, which can have an inferential interpretation, as we have seen earlier. What is special about the example treated there is the presence of the element of “astonishment” or of “change of knowledge”, which, in

45 Constat inattendu (Desclés & Guentchéva 2000: 101) or découverte inattendue (Guentchéva 1994: 19). 46 “Prenons l’exemple suivant : “Tiens, il a plu”. Le raisonnement abductif est ici le suivant : 1°) L’énonciateur sait que s’il pleut alors la chaussée est mouillée et, par ailleurs, son état de connaissance (ou de croyance) est : "il a fait beau aujourd’hui" ; 2°) l’énonciateur constate que la chaussée est mouillée ; 3°) par abduction il remonte à l’hypothèse ("il a plu") et doit alors réviser son état de connaissance d’où son énonciation qui signale son changement de connaissance. Il en résulte un effet d’étonnement, d’admiration…” (Desclés & Guentchéva 2000: 101).

Inferential Evidentiality and Abduction

45

this example, is made explicit by the particle tiens. What D&G claim is, crucially, that the “effect of astonishment” is not only concomitant with the abductive inferential process but actually results from it.47 In their example, it is however not clear why the effect of aston­ ishment should result from the inferential process (suggested by the perfect), rather than from the presence of the particle tiens. Inference and astonishment seem to be quite independent concepts; the attitude of astonishment may regard not only inferred information, but information obtained from all kinds of sources. In an example such as (13), which contains the mirative particle tiens as well, what clashes with the speaker’s previous expectations (the roadway is expected to be dry) is a directly observed fact (the roadway is wet) rather than an inferred content: (13) Tiens, la chaussée est mouillée! ‘Look at that, the roadway is wet!’

Example (14) in Guentchéva (2017: 544) is analogous to (13): (14)

Ax, to valjalo! Oh, it rain.l-​part.aor.impf.Neutr [‘Oh, it is raining!’]

We agree with the author that (14) “clearly express[es] a process happening at the time of speech but that does not match the enunciator’s expectations” (2017: 565) and “Clearly, in this case, there is no abduc­ tion nor deduction. […] There is simply a contradiction observed between what was expected and what actually happened” (2017: 565–​ 566). If, as shown by (13) and (14), speakers can express astonishment about non-​inferred information, it is not clear to us which exactly is the strong link between astonishment and abductive reasoning that allows the former to “be explained” by the latter.

47 Cf. also the formulation in Desclés & Guentchéva (2003: 50, our emphasis): “État de surprise, extension de l’état d’inference” (‘State of surprise, extension of the state of inference’).

46

Dendale and Miecznikowski

5.4. Our View on the Broad Version of D&G’s Claim In the broad version of D&G’s claim, abduction (and its associated property, plausibility) become defining features of the entire category of mediativity, rather than defining or characterizing elements of inferential evidentiality alone. This is confirmed by the following quote: We postulate here a precise definition of the mediative, which is grammaticalized or else expressed by lexical means. By mediative, we mean that which derives from an abductive reasoning in the sense of C. S. Peirce (1965) (Provôt & Desclés 2012: 47, our translation and emphasis).48

In our view, the broad conception of D&G’s claim is not tenable: it overstates the importance of inference for all types of mediativity/​ evidentiality, blurs the distinction between the three subcategories of mediativity/​evidentiality and hampers the endeavour undertaken in this article, i.e. a characterization of evidential inference. The broad version is also heavily dependent on the enunciative dimension the two authors underline in the notion of mediativity, a dimension that makes the notion of mediativity drift away from the mainstream understanding of the notion of evidentiality in terms of source of information, tracing back to Boas (1911: 124). We are quite convinced that D&G’s narrow claim, which relates abduction to inferential evidentials only, constitutes a useful starting point for an in-​depth reflection on the nature of inferential evidentials. In contrast, in our view the application of the notion of abduction to the analysis of all subcategories of mediative markers does not bring the discussion about inferentials any further. For this reason, as announced earlier, we will ignore the broad version of D&G’s claim in what follows and focus our attention on their narrow claim.

48 “Nous postulons ici une définition précise du médiatif, qui est grammatica­­ lisée [sic] ou bien exprimée [sic] par certains moyens lexicaux. Par médiatif, nous entendons ce qui relève d’un raisonnement abductif au sens de C. S. Peirce (1965)“. (Provôt & Desclés 2012: 47, our emphasis).

Inferential Evidentiality and Abduction

47

6. The Components of Abductive Reasoning according to D&G After having examined the two versions of D&G’s claim about the link between mediativity and abduction, we will now return to the properties of abduction, briefly sketched in section 3, and examine them in more detail. 6.1. Abduction as a Type of Inference or Reasoning and as a Form or Argument Abduction is a type of inference or reasoning (and a cognitive mechanism, Desclés 2018: 394) that differs both from deduction (includ­ ing “deduction with a probable consequent”, cf. section 4.3) and from induction. As a form of argument, it corresponds to what is sometimes called “inverted modus ponens” (Aliseda 2000: 172; Fisher 2001: 369; Mazur 2014: 25), “modus ponens turned backward” (Charniak and McDermott 1985, cited by Josephson & Josephson 1996: 5), “reverse inference” (Hanson 1960: 103), or “false modus ponens” (e.g. Rigotti & Greco 2019: 212). It has the form indicated in (10), which we repeat here: (15) q p⇒q plausibly p / ​p is plausible

The elements that compose the abductive form of argument are: − −

q: the observed element, which asks for an explanation, an explicative hypothesis (Desclés & Jackiewicz 2006) p ⇒ q: a general law or rule (Desclés & Guentchéva 2001: 105; Desclés & Jackiewicz 2006: 36), which states a conditional

48

−

Dendale and Miecznikowski

relation between the propositions p and q, taking the form of a material implication (Desclés & Guentchéva 2018: 241–​242) plausibly p / ​p is plausible: a plausible hypothesis. (Desclés 1989: 49, note 1)

In what follows, we will comment on these three elements. 6.2. q Abductive inference is usually triggered by an observation (Guentchéva 2017: 564; Desclés 2009: 41) of what is termed fact, situation, or –​underlining the semiotic dimension –​traces, clues, data: “observed facts” (Desclés & Guentchéva 2001: 105), an “observed situa­ tion” (Guentchéva 2017: 564), “observable empirical data” (Desclés 1989: 130), “observed traces or clues” (Guentchéva et al. 1994: 145; Desclés 1995: 10), “un faisceau constaté d’indices pertinents” (‘a bun­ dle of observed relevant clues’, Desclés 2013: 218) or, for short, “observ­ ables” (Desclés 1994: 22). The perceptual access to these observables guarantees their status as facts: the element q is qualified by Guentchéva (2017: 564; see also Desclés 1989: 36 n.1) as “observed and therefore true”. Alternatively, the triggering clues can be “known” to the speaker: “Abduction (or retroduction) is based on facts (observed or known)” (Desclés & Guentchéva 2018: 242). In other words, speakers can access the clues “cognitively”: “à partir d’indices, souvent perceptifs (mais aussi parfois cognitifs)” (Guentchéva 2003: 180) (‘based on clues, often perceptual (but sometimes also cognitive)’). They are the input and starting point of the abductive “process”, “procedure”, “inference”, “act”, “reasoning” or “reconstruction”. Desclés 2018 gives the following short overview of “types of clues”: Several types of clues (or sources of plausible knowledge) are susceptible of triggering abductive inference: (i) clues from sensory perception or cognitive observations; (ii) clues linked to rumor or hearsay; (iii) information not obtained directly but through intermediaries; (iv) observed states concomitant with states interpreted as states [sic] resulting from (plausible) occurrences of prior events. (Desclés 2018: 396)

Inferential Evidentiality and Abduction

49

The status of element q is double: it is both the observed fact that asks for an explanation –​triggering the entire abductive process –​and it is the clue (or one of the clues) for the to-​be-​found hypothesis. 6.3. p ⇒ q The second element in the abductive form of inference as described by D&G is a “conditional relation or material implication” (Desclés & Guentchéva 2018: 241). Its content corresponds to what the authors call: “une loi ou règle (générale)” (‘a (general) law or rule’,49 Desclés 1989: 36 fn. 1; Desclés & Guentchéva 2001: 108 fn.10, Guentchéva 2003: 180 fn. 6), “common knowledge” (Desclés & Guentchéva 2018: 242), “un savoir commun mémorisé” (“memorized common knowledge”, Desclés & Guentchéva 2001: 98). This knowledge has to be “shared” (Guentchéva 2017: 565) or “presumably shared” knowl­ edge (un savoir “(supposé) partagé”, Desclés 2009: 41, n.31; “un savoir présumé commun”, Desclés 2009: 41): clues only lead to abductive inference if there is shared knowledge (or knowledge construed and demonstrated within a theoretical framework) of a stable relation explicitly linking the formulation of the hypothesis in question to the observed clues. (Desclés 2018: 396)

It can, however, also be “un savoir local que l’énonciateur suppose partagé par son interlocuteur, dans une situation particulière d’énonciation” (‘local knowledge that the speaker presumes to be shared by the hearer, 49 With regard to the distinction between laws and rules, the authors comment: “Selon que la conditionnelle “p = > q” est une loi qui relève d’un savoir commun et partagé par tous les utilisateurs, ou qu’elle détermine une règle, qui relève d’un savoir plus local et moins bien partagé, le raisonnement (déductif ou abductif) qui utilisera la loi ou la règle, aura une portée plus ou moins universelle ou plus ou moins locale”. (Desclés & Guentchéva 2001:108– 109 fn. 10, our emphasis). ( ‘Depending on whether the conditional sentence “p = > q” is a law which relates to common knowledge and is shared by all users, or whether it determines a rule, which relates to a more local and less well shared knowledge, the reasoning (deductive or abductive) which will use the law or the rule, will have a more or less universal or more or less local scope’).

50

Dendale and Miecznikowski

in a specific utterance situation’) (Desclés & Guentchéva 2001: 108 fn.10, 115). In the law or rule, the elements put into relation, symbolized as p and q in (15), represent respectively a “hypothesis” (H), which refers to an event the speaker isn’t (or wasn’t) able to observe directly (Guentchéva 2011: 130), and an “observed fact” (C): let H and C be propositions that express, respectively, a hypothesis and an observed fact; we also state a law linking the hypothesis and the observation, expressed by the following logical implication: ‘H ⇒ C’; reasoning by abduction consists in noting the truth of C and in inferring the plausibility of the hypothesis H on that basis: C is thus considered as a clue for H. (Desclés 1989: 49, n. 1, our translation)50

As discussed in section 3, H and C are typically, albeit not necessarily, causally related; most of the examples given and analyzed by Desclés and Guentchéva illustrate that typical case. The law can be complex. Instead of stating a simple relation between H and C, the consequent of the material implication can refer to a conjunction of facts ensuing from the hypothesis: Abductive reasoning works best when, on the one hand, we have a more complex law like: ‘H ⇒ [C1 AND C2 AND… AND Cn]’, which predicts the finding of facts on the basis of the hypothesis H, and on the other hand, we verify a bundle of clues C1, C2, and Cn (observed facts); the plausibility of hypothesis H is then reinforced. It suffices that one of the clues C1 [C2…Cn] is false and hypothesis H will have to be rejected. (Desclés 1989: 49, n.1, our translation)51

50 “soient H et C des propositions qui expriment respectivement une hypothèse et un constat; soit une loi reliant hypothèse et constat exprimée par l’implication logique suivante : ‘H ⇒ C’ ; le raisonnement par abduction consiste à constater la vérité de C et à en déduire la plausibilité de l’hypothèse H : C’est ainsi considéré comme un indice pour H”. (Desclés 1989: 49, n. 1). 51 “Le raisonnement abductif fonctionne mieux lorsque d’une part, on a une loi plus complexe comme : ‘H ⇒ [C1 ET C2 ET … ET Cn]’, qui prédit le constat de faits à partir de l’hypothèse H, et d’autre part, on vérifie le faisceau d’indices C1, C2, et Cn (faits constatés), l’hypothèse H voit alors renforcer sa plausibilité ́. Il suffit que l’un des indices C1 soit faux pour entrainer le rejet de l’hypothèse H”. (Desclés 1989: 49, n.1).

Inferential Evidentiality and Abduction

51

As a premise in the abductive syllogism, the conditional relation p ⇒ q is considered to be true (Desclés & Guentchéva 2001: 108). It can have a probable consequent and display the following form: [p ⇒ probably (q)], meaning “when we have p, q is probable”. 6.4. (plausibly) p The third element of the abductive form of argument, p, corresponds to the antecedent of the “law” (p ⇒ q, H ⇒ C) and hence to the proposition put forward as a hypothesis by the speaker. It is said to be most often an explicative hypothesis, which typically orients the reasoning toward a plausible cause of what has been observed (Desclés & Jackiewicz 2006: 35). Abduction thus leads to hypotheses (Desclés 2016: 33). That is why Peirce, when distinguishing different inference types, used hypothesis as an alternative term for abduction itself (see section 3). An important property of an ‘abduced’ hypothesis is that it is plausible. This property distinguishes abduction both from deduction and induction. In D&G’s terms, deduction leads to a conclusion, rather than to a hypothesis, a conclusion that is necessarily true –​or, in some cases, probable (cf. section 4.3) –​, if and only if the premises are true (or, respectively, probable). Induction, by generalization, leads to a law that is most often probable. The authors insist on the fact that plausibility is a property tightly associated with abduction: “abductive reasoning leads to plausible hypotheses but not to probable conclusions” (Desclés & Guentchéva 2018: 246). What considerably varies in D&G’s formulations is the way in which the property of plausibility is linked to the abduced hypothesis: plausibility is said to be endorsed (“prise en charge”,52 Des­ clés 2009: 41), uttered (“énoncée”, Desclés & Guentchéva 2000: 99),53 52 The formula that represents this in D&G’s framework is: “DIS (plausible par abduction (ce qui est dit)) JE”, where DIS is the first-​person singular form of the present tense of dire ‘to say’. 53 Cf. “par abduction, l’énonciateur énonce la plausibilité de ce qui est dit” ‘by abduction, the enunciator utters the plausibility of what is said’. Note that the formulation “énoncer la plausibilité” seems to establish plausibility as an ‘object’ of the enunciation. Desclés (2003: 53) even considers the plausibility judgment a

52

Dendale and Miecznikowski

expressed (“exprimée”, Desclés 2003: 53), formulated (“formulée”, Desclés & Jackiewicz 2006: 35), stated (“posée”, Desclés & Jackiewicz 2006: 37), inferred (“inférée”, Desclés & Jackiewicz 2006: 37) and even deduced (“déduite”, Desclés 1989: 49). A hypothesis is considered plau­ sible (“considerée comme plausible”, Desclés & Guentchéva 2000: 98), or made plausible (“rendue plausible”, Guentchéva 2003: 180). But what exactly is plausibility? And how does it differ from probability? We did not find one single explicit definition of plausibility in D&G’s work. The notion is simply coupled to that of abduction, which, as a form of inference, produces plausibility in virtue of its characteristic logical form: The plausibility –​which is not a probability –​of the hypothesis established by abductive reasoning does not result from the uncertainty of the premises and therefore from the probability of the hypothesis considered, but from the very mode of reasoning, which, starting from the observed facts, does not produce certainty. (Desclés & Guentchéva 2001: 21, our translation)54

The authors indicate, however, what the effect of a plausible hypothesis is: “stating that a hypothesis is plausible leaves the field open for competing explanations (often yet unknown)”. (Desclés & Guentchéva 2018: 246). The plausibility of a hypothesis can be “strong” or “weak” (Desclés 2003: 53), depending on the number, the quality and the con­ gruency of the clues. Strong plausibility is assimilated by the authors with “probability” (“abductively probable”); weak plausibility with “mere possibility” (“abductively possible”):

meaning, when he mentions “l'énonciation de la signification "p est plausible" ”, ‘the enunciation of the meaning "p is plausible"’. This idea is quite surprising. It makes more sense to us to say that the object of an enunciation is the utterance (l’énoncé). The plausible character of the inferred utterance then, in our view, can be said to be ‘shown’ –​in the same way as the assertive character (or ‘force’, in the Anglo-​Saxon tradition) of an utterance is not ‘enunciated’, but ‘shown’. 54 “La plausibilité –​qui n'est pas une probabilité –​de l'hypothèse établie par le raisonnement abductif vient non pas de l'incertitude des prémisses et donc de la probabilité de l'hypothèse envisagée mais du mode même du raisonnement qui ne transmet pas la certitude à partir des faits constatés”. (Desclés & Guentchéva 2001: 21).

Inferential Evidentiality and Abduction

53

The proposition p has a strong plausibility when there is a bundle of very characteristic or congruent clues; in this case, p is probable. The proposition p has a weak plausibility when the clues are weak or few or not very characteristic; in this case, p is simply possible. As we have shown in (Desclés and Guentchéva, 2001), it is essential to clearly distinguish the notions of epistem­ ically POSSIBLE and PROBABLE from the notions of abductively POSSIBLE and PROBABLE; the latter are the traces of an underlying abductive reasoning. (Desclés 2003: 53, our translation)55

All this makes the authors speak of “plausibilité (possible ou probable)”56 (Desclés 2003: 53). In both cases, plausible implies “only plausible” (Guentchéva 2004: 26; Desclés 2016: 80) and thus uncertain (Guentchéva 2017: 565; Desclés 2018: 397) or rather non-​certain (Des­­ clés, 2003: 53). Abduction produces a form of non-​assertive enuncia­ tion; therefore, abductive markers do never commit the speaker to the truth of his/​her utterance (Desclés 2018: 397).

7. Discussion 7.1. Originality and Impact of D&G’s (Narrow) Claim D&G’s claim is original and unique. It differs from the predominant, “classical” way of describing evidential inference, which focuses on the type of evidence the inference is based on (clues, premises, laws…,

55 “La proposition p a une plausibilité forte, lorsqu'il y a un faisceau d'indices très caractéristiques ou congruents ; dans ce cas, p est probable. La proposition p a une plausibilité faible, lorsque les indices sont faibles ou peu nombreux ou peu caractéristiques ; dans ce cas, p est simplement possible. Comme nous l'avons montré dans (Desclés et Guentchéva, 2001), il est indispensable de bien dis­ tinguer les notions de POSSIBLE et de PROBABLE épistémiques des notions de POSSIBLE et de PROBABLE abductifs qui sont les traces d'un raisonnement abductif sous-​jacent”. (Desclés 2003: 53). 56 We can wonder whether there isn’t a contradictio in terminis in the combination “probable plausibility”, given the fact that the authors stress the fundamental difference between probability and plausibility.

54

Dendale and Miecznikowski

see section 1). D&G’s claim is also radical because, in its narrow version, it seems to apply without exception to all markers of inferential mediativity, and only to them. This means that if a particular language item expresses a form of inference that is not abductive, it cannot be an (inferential) evidential in their analysis, as we pointed out in section 4.3. Their claim thus offers a clear and firm answer to the question about the nature of “evidential inference” formulated in section 1. The authors’ sustained efforts to defend that original and radical claim have been the motivation for the present article. Their idea has not drawn the attention it deserves, in our view, and has not triggered much novel thinking about the nature of evidential inference. It was picked up –​as was the term mediativity –​by several scholars (especially authors contributing to Guentchéva 1996 and Guentchéva & Landaburu 2007), who however often limit themselves to just mentioning the term abduction, without theorizing the concept or using its full potential in their linguistic analysis (e.g. Roulet 1996, Donabédian 2001: 424, Barbet 2012, Gosselin 2015; an exception is Dendale & De Mulder 1996). In the Anglo-​Saxon tradition, D&G’s work remained largely unknown. We see three possible reasons for that. First, it may be the case that the notion of abduction –​debated very intensely in epistemology, logic and artificial intelligence nowadays –​ is not felt to be theoretically or empirically useful by linguists. According to D&G themselves, “Abduction is often misunderstood not only by epistemologists and philosophers of science but also by linguists” (Desclés & Guentchéva 2018: 244; see also Desclés & Jackiewicz 2006: 36). We think that, rather than only being “misunderstood”, the notion simply remained (surprisingly) unknown to linguists, especially to those working on evidentiality. The term abduction, together with its cognate terms (see section 7.2. below), rarely appear in work on evidentiality,57 except from some brief comments to the notion by Givón (1982: 25, 44–​45), two mentions of the term in Faller’s PhD (2002) and, most importantly, research performed by scholars in some way connected to Desclés & Guentchéva, in the fields of both linguistic typology and French linguistics (e.g. Kronning 2004, Celle 2008, Rossari, Cojocariu, Ricci & Spiridon 2009, Azzopardi 2011, Vetters 2012, Messi 2019). 57 No mention of the notion is made in Aikhenvald’s 2004 monograph.

Inferential Evidentiality and Abduction

55

Secondly, most of D&G’s work on the notion, until recently, was published in French, especially the theoretically most important studies (Desclés 1996, Desclés & Guentchéva 2001, Desclés & Jack­­­ iewicz 2006 and Desclés & Guentchéva 2013); exceptions are Desclés & Guentchéva 2018 and Desclés 2018, 2000.58 Thirdly, the authors’ deliberate choice to stick to the term ­mediativity59 rather than shifting to the nowadays internationally more established term evidentiality60 –​plus the fact that D&G prefer to highlight, in several studies, the differences between the notions of mediativity and evidentiality rather than their areas of overlap (cf. the following citations), may have played a role: I tried to show that the notion of evidentiality is not to be confused with that of mediation (Guentchéva 2004: 30, our translation and emphasis).61 The semantic value of the utterance (1) [Il serait donc parti ‘So, he must have left //​He must have left, then’] is called ‘mediative’ rather than ‘evidential. (Desclés and Guentchéva 2018: 27 endnote 2, our emphasis)

D&G have, of course, theoretical reasons to keep mediativity and evidentiality separated (see section 2), but the consequence may have been a shortfall in theoretical impact of their claim on inferential evidentiality in the field of research on evidentiality. Despite the limited impact their claim has had until now, we believe (contra our own early study, Dendale & De Mulder 1996) that, in its narrow version, it offers a promising way of looking at evidential inference and inferential evidentiality. Compared to the characterization 58 Among D&G’s writings, we only counted ten studies in total on the topic written in English. 59 As did other influential scholars in French linguistics not belonging to D&G’s research group (e.g. Kronning 2004, Anscombre, Oppermann-​ Marsaux & Somolinos 2014). Anscombre also imported the term into Spanish as mediatividad (Anscombre 2008, 2014). 60 The terminological preeminence of the term evidentiality also appears from its calques in many European languages: e.g. evidenzialità, evidencialidad, evidențialitate, Evidentialität, evidentialiteit, evidentialitet , etc. 61 “J’ai essayé de montrer que la notion d’evidentialité ne se confond pas avec celle de médiation”. (Guentchéva 2004: 30, our emphasis). See also the citation at p.25, and the whole article of Guentchéva (2014a).

56

Dendale and Miecznikowski

of the latter notions in terms of types of evidence (i.e. types of contextually salient premises, cf. section 1), it has the potential of providing a more fine-​grained description of the properties of evidential inference and thus of the properties of evidential inference markers. This is true especially for the marking of the type of inference called Inferred (as opposed to the one called Assumed) by Aikhenvald (2004). D&G’s presentation of abductive inference schemes reveals that the salient premises the typological tradition focuses on –​sensory, perceptual, circumstantial evidence, etc. –​correspond only to a part of an utterance’s information source, namely to the “observed traces or clues” (Guentchéva et al. 1994: 145; Desclés 1995: 10, cf. section 6.2. above), and that these have to be complemented by “a (general) law or rule” (Desclés 1989: 36 fn. 1; Guentchéva 2003: 180 fn. 6, cf. section 6.3.) to allow the formulation of the plausible conclusion. No conclusion is indeed drawn on the basis of “observations” alone. In other words, D&G’s work reminds us that inferential procedures derive conclusions from configurations of premises, typically minor premises (the propositional representation of specific facts, e.g. “observed traces and clues”) and major premises (the more general “conditional laws” or “material implications”). It invites linguists to not exclusively focus their attention on the premises that are most salient in context,62 but on all premises that are required to derive a conclusion and on the overall logical form of the inferential scheme indicated by an inferential mediative/​evidential marker. 62 In the case of markers of the type Inferred, the evidence considered by lin­ guists usually corresponds to a minor premise. In the case of markers of the type Assumed, major premises have attracted more scholarly attention, as shows for instance the definition of these markers in Aikhenvald’s (2004) glos­ sary: “ASSUMED EVIDENTIAL: information source based on conclusions drawn on the basis of logical conclusion and general knowledge and experience” (Aikhenvald 2004: 391, our emphasis). General knowledge and (repeated) experience are indeed more likely to generate ‘laws’, i.e. major premises, than to provide the speaker with knowledge about specific facts, i.e. minor premises. The indistinct category of evidence –​that on which conclusions are ‘based’ (see Aikhenvald’s quote) –​, does not, however, allow to clearly determine which kind of premises are at stake and how they must be related to each other to derive acceptable conclusions. This and other points will be the object of a further papers we are currently preparing.

Inferential Evidentiality and Abduction

57

This view has clear affinities with the way inference in discourse is seen by theoreticians of argumentation, who use inference schemes to describe the configuration of premises underlying arguments (van Eemeren & Grootendorst 1992, Walton, Reed & Macagno 2008, Rig­­ otti & Greco 2019). Recent work suggests that inference schemes can usefully be employed also in the linguistic description of inferential evidential markers and constructions (Rocci 2017, Miecznikowski & Musi 2015, Miecznikowski 2020) and thus echoes D&G’s exhortation to reconstruct the inferential procedure signaled by those markers and constructions more completely than what is commonly proposed in the linguistic literature on evidentiality. 7.2. Issues, Unanswered Questions and Untreated Data in D&G’s Proposal As argued in section 7.1., D&G’s claim is a good starting point for a close examination of the properties of inferential evidentials. There is a lot of work ahead, however, for D&G’s presentation of the notion of abduction as definitional of inferential evidentials raises a series of questions. Continuing the argument started in this paper, we will, in further work, compare the properties of abductive inference as they emerge from D&G’s writings (see section 6) with those presented in the vast literature on abduction (Fann 1970, Aliseda 2006), retroduc­­ tion (Remnant 1948) or hypothesis (Peirce 1931,1935), conjecture (Ma & Pietarinen 2018), abductive guess (Andersen 2017), presumption (Rescher 2006), presumptive inference or presumptive reasoning (Wal­­ ton 2001), inference to the best explanation (Harman 1965), plausible reasoning (Rescher 1976, Pólya 1989), heuristic reasoning or heuristic syllogism (Pólya 1989), defeasible reasoning (Pollock 1987), default reasoning (Reiter 1978, Etherington & Crawford 1999) and nonmono­­ tonic reasoning or inference (Moore 1985, Reiter 1987, Brewka, Dix & Konolige 1997), etc. We will limit ourselves here to four points that call for further investigation. A first issue concerns the difference between plausibility and probability. The second notion refers to the result of an epistemic-​ modal evaluation, as it is commonly defined in the literature, whereas

58

Dendale and Miecznikowski

the first (hardly used in the literature on evidentiality) refers to a property of the conclusion of an abductive inference. Albeit important in D&G’s work, the two notions, which both imply uncertainty, are not defined contrastively by the authors. They are even sometimes equated with each other, as we saw in section 6.4. Presenting plausibility as “plus vague, moins décisive, moins complètement exprimée” (‘vaguer, less decisive and less completely expressed’) than probability (Pólya 1989: 108–​109, cited in Desclés & Guentchéva 2001: 111) is far not enough to grasp the defining properties of plausi­ bility and to distinguish plausibility judgments from probability judgments. Besides other studies (Pólya 1949, Hamblin 1959), Rescher’s monograph (1976) on plausible reasoning could contribute to clarify that point. A second series of issues concerns D&G’s conceptualization of the “conditional law” (or major premise) in abductive reasoning. First, the authors heavily stress (cf. section 6.3.) that the law has to represent “(presumably) shared” or “memorized” information. This is by no means a generally acknowledged requirement, however. Andersen (1973: 775) observes that “the law to which an abductive inference appeals may be an already established truth; or it may be a new, tentative generalization”. Laws that do not yet have the status of shared knowledge seem particularly important when abduction takes the form of a creative process in scientific reasoning (Kapitan 1990, Desclés 1996, 2000, Paavola 2005). As Bonfantini and Proni (1983: 133) put it: […] the novelty of the abduction is still sharper and stronger when the principle expressed in the major premise is a new theoretical law rather than a universally accepted scientific law. In that case the abductive conclusion is "a new idea" in absolute terms: it is not just the application of the general principle to the subject of the investigation that is new, it is the principle that is new too. Hence the conclusion was not even potentially included in the existing store of knowledge. For an example of the latter type of abduction, which is the most fertile in scientific research, one could usefully take the process of reasoning with which Bohr interpreted the mystery of the gaps in the lines of the hydrogen spectrum (cf. Bonfantini and Macciò 1977: 88–​102).

Abducing hypotheses can be seen as searching not only for a hypothesis but also for the missing information in the protasis of a new

Inferential Evidentiality and Abduction

59

conditional law –​representable as: [? =​> C].63 Formulating a hypothesis then amounts to formulating the corresponding law at the same time: While it may be true to say that we first come to be acquainted with A [the Hypothesis] “in the course of making an inference” to A’s plausibility –​that is, we might never have come upon A without seeking a plausible explanation for C –​it does not follow that the first acquaintance with A is the result of an inferential transition. […] while the conditional premise containing A is logically prior to the conclusion, this does not mean that it is arrived at temporally prior to the conclusion, for the premise and conclusion may be “simultaneously arrived at”. (Kapitan 1990: 502).

At a descriptive level, the question arises of how important given, shared laws actually are in evidential inference. Second, as we underlined repeatedly, the conditional relations in the abductive inference scheme are almost always considered by D&G to be causal laws. Although the authors mention the possibility of non-​causal links, e.g. “praxeological rules” (see section 3), they neither explore those in detail in their examples, nor explain why they so strongly focus on causal reasoning. Dendale’s 1991 study found that inferential markers can contribute to realize various conceptual operations, including operations without any causal reference such as approximate calculations or categorizations. Taxonomies of inference schemes and loci underlying argumentation (van Eemeren & Grooten­ dorst 1992, Walton, Reed & Macagno 2008, Rocci 2017, Rigotti & Greco 2019), too, suggest that inferences can be based on a quite wide range of conceptual relations between the premises and the conclusion, e.g. causality, concomitance, definition, classification, analogy or authoritative source. Can the concept of abductive inference be extended to include various schemes based on non-​causal relations or is it actually confined to primarily causal relations for some theoretical reason? Third, there is a problem with the validity of the laws or conditional relations D&G formulate. Most of these are not universal laws, but default generalizations or generalized statements entering forms of default reasoning (Reiter 1978, 1980, 1987). A law like “s’il pleut la route est mouillée” (‘if it rains the road is wet’) (Desclés & Guentchéva 63 Cf. McAuliffe: “abduction poses a hypothesis as a question” (2015: 303).

60

Dendale and Miecznikowski

2000: 101) will not be valid for roads that are covered (in tunnels or under viaducts) and a law like “Si on est malade alors on a mauvaise mine” (‘If we’re sick then we look bad’) is clearly not a universally valid conditional relation. To account for this, D&G add elements that limit the universal validity in some of their laws, e.g. the adverb souvent “often” in “Quand nous approchons d'une terre, nous voyons souvent des oiseaux” (‘When we approach land we often see birds’) (Desclés & Guentchéva 2000: 110). In ordinary life most laws are indeed not universally valid. That limited validity seems to be acknowledged by D&G in what they call “deduction of a probable consequence” (Desclés & Guentchéva 2000, 2018): e.g. “Si quelqu'un travaille beaucoup alors il est probablement fatigué” (‘If someone works a lot s/​he is probably tired’). It is also relevant for the terminological distinction they make between “laws” and “rules” (see fn. 49). In their abductive schemes, in contrast, the limited validity of the law is not given a theoretical status. We think that the relation between abduction and default reasoning would deserve closer attention. To answer these various open questions, it seems necessary to widen the scope of research to include nonmonotonic forms of inference, which all have the characteristics to be plausible and defeasible (see section 5.2.), just like abduction, but differ from abduction in other respects.64 The literature on forms of nonmonotonic inference is rich and profuse. D&G, mainly inspired by Peirce and Pólya, neither explicitly link abduction to nonmonotonic inference in their writings, nor examine other forms of nonmonotonic reasoning besides abduction. In our opinion, that narrow focus on abduction must be abandoned in order to fruitfully develop the useful suggestions that emerge from D&G’s work concerning the law, or major premise, of evidential inference schemes. A third issue, related to the previous one, is D&G’ split analysis of the meaning of “(weak) epistemic” devoir. It is generally recognized that the weak epistemic use of devoir differs both from the verb’s deontic use and from what Kronning (e.g.1996, 2001) describes as its 64 “The province of nonmonotonic reasoning is the derivation of plausible but not infallible conclusions from a knowledge base viewed abstractly as a set of formulas in a suitable logic” (Reiter 1987: 148).

Inferential Evidentiality and Abduction

61

“alethic” use (and which on the level of commitment corresponds to a “strong epistemic” use). According to D&G, depending on the content of the sentence in which it appears, weak epistemic devoir can either be an evidential marker of abductive inference or a marker of weak epistemic modality that encodes a “deduction based on a law with probable consequent” (see section 4.3). This analysis makes use of “probabilized” laws or rules, which are not clearly defined by D&G. It also splits weak epistemic devoir into two groups of uses, one belonging to the domain of modality, the other to the domain of evidentiality. We see no linguistic evidence (in the conditions of use, distribution etc. of evidential markers) to justify this attribution of uses to two distinct functional domains. The semantic differences that have been pointed out are mainly related to the direction of causality that is relevant in the speaker’s reasoning as it can be reconstructed from the content of the utterance in context: from cause to consequence (epistemic modality for D&G) vs. from consequence to cause (abductive inferential evidentiality for them). Since both types of reasoning derive new information from other pieces of information, both uses of weak epistemic devoir should be considered evidential inferentials. A fourth issue concerns the definition of the notion of “Hypothesis” in D&G’s abductive inference scheme. Symbolized as H, this element appears both as the protasis of the conditional law (H =​> C) and as the conclusion of the abductive inference schema. We also observed that Peirce equates abductive inference (as a whole) to “hypothesis”, probably by metonymy. Hanson (1960: 101) qualifies a hypothesis “as a natural law”, which would mean that Hypothesis could refer to the law as a whole. So, what exactly is the referent of Hypothesis? Can an evidential marker of inference be called a “marker of hypothesis”? This point also needs further investigation.

8. Conclusion The question at the outset of this article regarded the nature of evidential inference”, viz. the inference inferential evidentials refer to. Two

62

Dendale and Miecznikowski

common strategies used to answer that question are: a characterization of the type of evidence (premises) the inference is based on and a characterization of the type of operation inferential evidentials refer to. In the present article we explored the second strategy, discussing the work of two French linguists, Jean-​Pierre Desclés and Zlatka Guentchéva. These authors defend the idea that evidential inference is fundamentally abductive in nature, referring to the notion of abduction introduced by Charles Sanders Peirce besides those of deduction and induction. We distinguished and briefly described what we call D&G’s “narrow claim” and “broad claim” about the nature of evidential inference and then focused on the narrow claim, the one among the two we found useful to investigate inferential evidentials. For D&G, inferential markers of evidentiality (or mediativity as they term it) always point to an abductive inference schema, comprising three elements: an observation, a general law and a hypothesis concluded from these premises. We described the properties of these elements according to the writings of D&G and pointed out a series of issues and unanswered questions raised by their analysis and characterization of abduction. In research in progress, we intend, first, to critically assess the other –​more common, but in certain respects problematic –​strategy followed by linguists to answer the question at the outset: the strategy that distinguishes between sensory evidence and non-​ sensory (conceptual) evidence and leads to oppositions such as Inferred versus Assumed (Aikhenvald 2004). We will further pursue our line of reason­ ing by relating D&G’s view to work by other authors who have analyzed abduction and related concepts –​not only in the realm of linguistics but also in (informal) logic, epistemology, artificial intelligence and argumentation theory. Our final aim is to give a comprehensive and structured overview of the properties of evidential inference, which could prove useful in describing inferential evidentials and the fine semantic and pragmatic differentiations that can be observed within this particularly rich category of evidential markers.

Inferential Evidentiality and Abduction

63

References Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: Oxford University Press. Aliseda, Atocha 2000. Abduction as epistemic change: a Peircean model in artificial intelligence. In Flach, Peter A. /​Kakas, Antonis C. (eds) Abduction and Induction: Essays on Their Relation and Integration, Dordrecht: Kluwer Academic Press, 45–​48. Aliseda, Atocha 2006. Abductive Reasoning. Dordrecht: Springer. Andersen, Henning 1973. Abductive and Deductive Change. Language: Journal of the Linguistic Society of America 49/​4, 765–​ 793. Andersen, Henning 2017. Abduction. In Ledgeway, Adam /​Roberts, Ian (eds) The Cambridge Handbook of Historical Syntax, Cambridge: Cambridge University Press, 301–​321. Anscombre, Jean-​Claude 2008. La polifonía: nociones y problemas. Archivum (Oviedo) 58–​59, 21–​51. Anscombre, Jean-​Claude 2014. Mediatividad, Omni-​locutores, y marcadores del discurso. El caso del francés actual: on sait que, on dit que, on prétend que, on raconte que, on admet que. Archivum (Oviedo) 64, 7–​26. Anscombre, Jean-​ Claude /​Oppermann-​ Marsaux, Evelyne /​Rodríguez Somolinos, Amalia (eds) 2014. Médiativité, polyphonie et modalité en français: études synchroniques et diachroniques. Paris: Presses Sorbonne nouvelle. Azzopardi, Sophie 2011. Le Futur et le Conditionnel: valeur en langue et effets de sens en discours. Analyse contrastive espagnol /​ français. Université de Montpellier. Bally, Charles 1965 [1932]. Linguistique générale et linguistique française. Bern: Francke. Barbet, Cécile 2012. Devoir et pouvoir, des marqueurs modaux ou évidentiels? Langue française 173/​1, 49–​63. Benveniste, Émile 1966. Problèmes de linguistique générale. Paris: Gallimard. Boas, Franz 1911. The Mind of Primitive Man: A Course of Lectures Delivered before the Lowell Institute, Boston, Mass., and the

64

Dendale and Miecznikowski

National University of Mexico, 1910–​1911. New York (State): ­Macmillan. Bonfantini, Massimo A. /​Macciò, Marco 1977. La neutralità impossibile. Quantistica, relatività, evoluzione biologica. Milano: Mazzotta. Bonfantini, Massimo A. /​Proni, Giampaolo 1983. To Guess or Not To Guess. In Eco, Umberto /​Sebeok, Thomas A. (ed) The Sign of the Three –​Dupin, Holmes, Peirce, Bloomington and Indianapolis: Indiana University Press, 119–​134. Brewka, Gerhard /​Dix, Jurgen /​Konolige, Kurt 1997. Nonmonotonic Reasoning. Stanford, California: Center for the Study of Language and Information. Caffi, Claudia 2007. Mitigation. Amsterdam, London etc.: Elsevier. Celle, Agnès 2008. Tense, modality and commitment in modes of mixed enunciation. Belgian Journal of Linguistics 22/​1, 15–​36. Charniak, Eugene /​McDermott, Drew 1985. Introduction to Artificial Intelligence. Menlo Park (CA): Addison-​Wesley. Chomsky, Noam A. 1972. Language and Mind (enlarged edn.). New York: Harcourt Press. Cornillie, Bert 2009. Evidentiality and epistemic modality: On the close relationship between two different categories. Functions of Language 16/​1, 44–​62. Crothers, Edward J. 1978. Inference and coherence. Discourse Processes 1/​1, 51–​71. Culioli, Antoine 1990/​1999. Pour une linguistique de l'énonciation, tomes 1, 2, 3. Paris: Ophrys. DeLancey, Scott 1997. Mirativity: The grammatical marking of unexpected information. Linguistic Typology 1, 33–​52. DeLancey, Scott 2001. The mirative and evidentiality. Journal of Pragmatics 33/​3, 369–​382. DeLancey, Scott 2012. Still mirative after all these years. Linguistic Typology 16, 529–​564. Dendale, Patrick 1991. Le marquage épistémique de l'énoncé. Esquisse d'une théorie avec applications au français. Unpublished PhD thesis. University of Antwerp. Dendale, Patrick 1993. Le conditionnel de l'information incertaine: marqueur modal ou marqueur évidentiel? In Hilty, Gerold (ed) Actes

Inferential Evidentiality and Abduction

65

du XXe Congrès International de Linguistique et Philologie Romanes, Université de Zurich, vol. 1. Tübingen: Francke. Dendale, Patrick 2018. Évidentialité ou non-​prise en charge ? Le cas du conditionnel épistémique en français. Une réanalyse. Langue française 200/​4, 63–​76. Dendale, Patrick 2020. Are "Modal Adverbs" automatically Modal Markers? The case of French certainement with its epistemico-​ modal and its evidential use, Anuari de filologia. Estudis de lingüística 10, 39–​76. Dendale, Patrick 2022. Evidential markers in French. In Wiemer, Björn /​ Marín Arrese, Juana I. (eds) Evidential Marking in European Languages. Toward a Unitary Comparative Account. Berlin: De Gruyter, Akademie Verlag, 76–​114. Dendale, Patrick /​De Mulder, Walter 1996. Déduction ou abduction: le cas de devoir inférentiel. In Guentchéva, Zlatka (ed) L’énonciation médiatisée, Louvain and Paris: Peeters, 305–​318. Dendale, Patrick /​Tasmowski, Liliane 2001. Introduction: Evidentiality and related notions. Journal of Pragmatics 33/​3, 339–​348. Desclés, Jean-​Pierre 1976. Quelques opérations énonciatives. Logique et niveaux d’analyse linguistique. Paris: Klincksieck, 213–​242. Desclés, Jean-​Pierre 1989. Catégories grammaticales et opérations cognitives. Histoire, épistémologie, langage 11/​1, 33–​53. Desclés, Jean-​Pierre 1994. Réflexions sur les grammaires cognitives. Modèles linguistiques 23, 1–​27. Desclés, Jean-​Pierre 1995. Langues, Langage et Cognition: quelques réflexions préliminaires. ARBA 3, 1–​32. Desclés, Jean-​Pierre 1996. L’abduction, procédé d’explication en linguistique. Modèles linguistiques 16/​2, 33–​62. Desclés, Jean-​Pierre 2000. Abduction and non-​observability. Some examples from language, science and cognitive science. In Agazzi, Evandro /​Pauri, Massimo (eds) The Reality of the Unobservable, Dordrecht: Springer, 87–​112. Desclés, Jean-​Pierre 2003. Interactions entre les valeurs de pouvoir, vouloir, devoir. In Birkelund, Merete /​Kiærsgaard, Poul S. /​ Boysen, Gerhard (eds) Aspects de la modalité, Berlin/​Boston: De Gruyter, 49–​66. Desclés, Jean-​Pierre 2009. Prise en charge, engagement et désengagement. Langue française 162, 29–​53.

66

Dendale and Miecznikowski

Desclés, Jean-​Pierre 2013. Intersémiotique et langues naturelles. Signata 4/​4, 175–​226. Desclés, Jean-​Pierre 2016. A cognitive and conceptual approach to tense and aspect markers. In Guentchéva, Zlatka (ed) Aspectuality and Temporality. Descriptive and Theoretical Issues, Amsterdam: Benjamins, 27–​60. Desclés, Jean-​Pierre 2018. Epistemic modality and evidentiality from an enunciative perspective. In Guentchéva, Zlatka (ed) Epistemic Modalities and Evidentiality in Cross-​Linguistic Perspective, Berlin/​Boston: De Gruyter, 383–​402. Desclés, Jean-​Pierre /​Guentchéva, Zlatka 2000. Énonciateur, locuteur, médiateur dans l’activité dialogique. In Monod-​Becquelin, Aurore /​ Erikson, Philippe (eds) Les rituels du dialogue, Nanterre: Société d'éthnologie. Desclés, Jean-​Pierre /​Guentchéva, Zlatka 2001. La Notion d'abduction et le verbe devoir 'épistémique'. In Dendale, Patrick /​Van der Auwera, Johan (eds) Les Verbes modaux (=​Cahiers Chronos 8), Amsterdam/​Atlanta: Rodopi, 103–​122. Desclés, Jean-​Pierre /​Guentchéva, Zlatka 2003. Comment déterminer les significations du passé composé par une exploration contextuelle? Langue française 138. 48–​60. Desclés, Jean-​Pierre /​Guentchéva, Zlatka 2012. Universals and typology. In Binnick, Robert I. (ed) The Oxford Handbook of Tense and Aspect, Oxford: Oxford University Press, 123–​154. Desclés, Jean-​ Pierre /​Guentchéva, Zlatka 2013. L’abduction dans l’analyse sémantique. In Norén, Coco /​Jonasson, Kerstin /​ Nølke, Henning /​Svensson, Maria (eds) Modalité, évidentialité et autres friandises langagières. Mélanges offerts à Hans Kronning à l’occasion de ses soixante ans, Bern: Peter Lang, 81–​102. Desclés, Jean-​Pierre /​Guentchéva, Zlatka 2018. Inference processes expressed by languages: deduction of a probable consequent vs. abduction. In Arigne, Viviane /​Rocq-​Migette, Christiane (eds) Theorization and Representations in Linguistics, Newcastle upon Tyne: Cambridge Scholars Publishing, 241–​266. Desclés, Jean-​Pierre /​Jackiewicz, Agata 2006. Abduction et prise en charge énonciative de la causalité. Linx 54, 35–​47. Deutscher, Guy 2002. On the misuse of the notion of ‘abduction’ in linguistics. Journal of Linguistics 38/​3, 469–​485.

Inferential Evidentiality and Abduction

67

Diewald, Gabriele /​Smirnova, Elena 2010. Evidentiality in German: Linguistic Realization and Regularities in Grammaticalization, Berlin/​New York: Mouton de Gruyter. Donabédian, Anaid 2001. Towards a semasiological account of evidentials: An enunciative approach of -​er in Modern Western Armenian. Journal of Pragmatics 33/​3, 421–​442. Etherington, David W. /​Crawford, James M. 1999. Toward efficient default reasoning. Logical Foundations for Cognitive Agents, Berlin/​Heidelberg: Springer, 137–​150. Faller, Martina 2002. Semantics and Pragmatics of Evidentials in Cuzco-​Quechua. Stanford University. Fann, Kuang T. 1970. Peirce's Theory of Abduction. The Hague: Nijhoff. Fischer, Hans R. 2001. Abductive reasoning as a way of worldmaking. Foundations of Science 6/​4, 361–​383. Frawley, William 1992. Linguistic Semantics. Hillsdale, NJ: Lawrence Erlbaum. Givón, Talmy, 1982. Evidentiality and epistemic space. Studies in Language: International Journal Sponsored by the Foundation 'Foundations of Language' 6/​1, 23–​49. Gosselin, Laurent 2015. De l'opposition modus /​dictum à la distinction entre modalités extrinsèques et modalités intrinsèques. Bulletin de la Société de linguistique de Paris 110/​1, 1–​50. Guentchéva, Zlatka 1985. Contributions à l'étude des catégories grammaticales en bulgare littéraire contemporain. Université de Paris VII. Guentchéva, Zlatka 1990. Enonciation médiatisée en bulgare. Revue des études slaves 62/​1–​2, 179–​196. Guentchéva, Zlatka 1994. Manifestations de la catégorie du médiatif dans les temps du français. Langue française 102/​1, 8–​23. Guentchéva, Zlatka 2003. Degrés de distanciation énonciative. In Quattara, Aboubakar (ed) Parcours énonciatifs et parcours interprétatifs. Théories et applications, Paris: Ophrys, 171–​183. Guentchéva, Zlatka 2004. La notion de médiation dans la diversité des langues. Les médiations langagières 1, 11–​33. Guentchéva, Zlatka 2011. L'opération de prise en charge et la notion de médiativité. In Dendale, Patrick /​Coltier, Danielle (eds) La prise en charge énonciative, Bruxelles: De Boeck/​Duculot, 117–​142.

68

Dendale and Miecznikowski

Guentchéva, Zlatka 2013. Médiativité ou ‘evidentiality’? Alternance de référentiels temporels dans la narration en bulgare. Contrastive Linguistics 38/​2–​3, 83–​109. Guentchéva, Zlatka 2014a. Aperçu des notions d’évidentialité et de médiativité. Cahiers de lexicologie 105, 57–​77. Guentchéva, Zlatka 2014b. Peut-​on identifier, et comment, les marqueurs dits ‘médiatifs’? In Anscombre, Jean-​ Claude /​Oppermann-​ Marsaux, Evelyne /​Rodriguez Somolinos, Amalia (eds) Médiativité, polyphonie et modalité en français : études synchroniques et diachroniques, Paris: Presses Sorbonne Nouvelle, 35–​50. Guentchéva, Zlatka 2017. An enunciative account of admirativity in Bulgarian. Review of Cognitive Linguistics 15/​2, 540–​575. Guentchéva, Zlatka ed. 1996. L'énonciation médiatisée. Leuven: Éditions Peeters. Guentchéva, Zlatka /​Landaburu, Jon (eds) 2007. L'Enonciation médiatisée II: Le Traitement épistémologique de l'information: Illustrations amérindiennes et caucasiennes. Leuven, Paris: Peeters. Hamblin, Charles L. 1959. The Modal "Probably". Mind 68/​270, 234–​240. Hanson, Norwood R. 1960. Is there a logic of scientific discovery? Australasian Journal of Philosophy 38/​2, 91–​106. Harman, Gilbert H. 1965. The inference to the best explanation. The Philosophical Review 74/​1, 88–​95. Hastings, Arthur C. 1962. A Reformulation of the Modes of Reasoning in Argumentation. Ph.D. Dissertation, Northwestern University, Evanston, Illinois. Johanson, Lars 2018. Turkic indirectivity. In Aikhenvald, Alexandra Y. (ed) The Oxford Handbook of Evidentiality. 1 edn. Oxford: Oxford University Press, 510–​524. Johanson, Lars /​Utas, Bo (eds). 2000. Evidentials. Turkic, Iranian and Neighbouring Languages. Berlin/​New York: Mouton de Gruyter. Josephson, John R. /​Josephson, Susan G., 1996. Abductive Inference. Computation, Philosophy, Technology. Cambridge: Cambridge University Press. Kapitan, Tomis 1990. In what way is abductive inference creative? Transactions of the Charles S. Peirce Society 26/​4, 499–​512.

Inferential Evidentiality and Abduction

69

Kintsch, Walter /​Van Dijk, Teun A. 1978. Toward a model of text comprehension and production. Psychological Review 85/​5, 363–​394. Kronning, Hans 1996. Modalité, cognition et polysémie: Sémantique du verbe modal devoir. Uppsala: Acta Universitatis Upsaliensis. Kronning, Hans 2001. Pour une tripartition des emplois du modal devoir. In Dendale, Patrick /​Van der Auwera, Johan (eds) Les Verbes modaux (=​ Cahiers Chronos 8), Amsterdam/​ Atlanta: Rodopi, 67–​84. Kronning, Hans 2004. Modalité et médiation épistémiques. In Delamotte-​Legrand, Régine (ed) Les médiations langagières. Vol I: Des faits de langue aux discours, Rouen: Publications de l'Université de Rouen, 34–​65. Lazard, Gilbert 1956. Caractères distinctifs de la langue tadjik. Bulletin de la Société de Linguistique de Paris 51/​1, 117–​186. Lazard, Gilbert 1996, Le Médiatif en persan. In Guentchéva, Zlatka (ed) L'Énonciation médiatisée, Louvain/​Paris: Peeters, 21–​30. Lazard, Gilbert 1999. Mirativity, evidentiality, mediativity, or other? Linguistic Typology 3/​1, 91–​109. Lazard, Gilbert 2000. Le médiatif: considérations théoriques et application à l’iranien. In Johanson, Lars /​Utas, Bo (eds) Evidentials. Turkic, Iranian and Neighbouring Languages, Berlin/​Boston: De Gruyter, 209–​228. Lazard, Gilbert 2001. On the grammaticalization of evidentiality. Journal of Pragmatics 33/​3, 359–​367. Ma, Minghui /​Pietarinen, Ahti-​ Veikko 2018. Let us investigate! Dynamic conjecture-​making as the formal logic of abduction. Journal of Philosophical Logic 47/​6, 913–​945. Marín-​Arrese, Juana I., 2013. Stancetaking and inter/​subjectivity in the Iraq inquiry: Blair vs. Brown. In Marín Arrese, Juana I. /​Carretero, Marta /​Arús Hita, Jorge /​Van der Auwera, Johan (eds) English Modality: Core, Periphery and Evidentiality, Berlin: de Gruyter Mouton, 411–​445. Marín-​Arrese, Juana I., 2017. Multifunctionality of evidential expressions in discourse domains and genres: Evidence from cross-​ linguistic case studies. In Marín Arrese, Juana I. /​Haβler, Gerda /​ Carretero, Marta (eds.) Evidentiality Revisited: Cognitive

70

Dendale and Miecznikowski

Grammar, Functional and Discourse-​Pragmatic Perspectives, Amsterdam: John Benjamins, 195–​222. Mazur, Barry 2014. Is it plausible? The Mathematical Intelligencer 36/​1, 24–​33. McAuliffe, William H.B. 2015. How did abduction get confused with inference to the best explanation? Transactions of the Charles Sanders Peirce Society 51/​3, 300–​319. Messi, Hugues U.P. 2019. Les sources du savoir : l’expression de l’inférence en Français : une analyse sémantico-​pragmatique et rhétorico-​syntaxique. Universität Potsdam. Miecznikowski, Johanna 2020. At the juncture between evidentiality and argumentation. Journal of Argumentation in Context 9/​1, 42–​68. Miecznikowski, Johanna 2015. Inferential connectives: the example of Italian come si vede. Nouveaux Cahiers de Linguistique Française 32, 103–​118. Miecznikowski, Johanna /​Musi, Elena 2015. Verbs of appearance and argument schemes: Italian sembrare as an argumentative indicator. In Van Eemeren, Frans H. /​Garssen, Bart (eds) Reflections on Theoretical Issues in Argumentation Theory, Amsterdam: Springer, 259–​278. Moore, Robert C. 1985. Semantical considerations on nonmonotonic logic. Artificial Intelligence 25/​1, 75–​94. Oswald, Steve /​Greco, Sara /​Miecznikowski, Johanna /​Pollaroli, Chiara /​Rocci, Andrea (eds) 2020. Argumentation and Meaning. Semantic and Pragmatic Reflexions. (Special issue of the Journal of Argumentation in Context 9/​1). Paavola, Sami 2005. Peircean abduction: instinct or inference? Semiotica 153/​1–​4, 131–​154. Palmer, Frank R. 1986. Mood and Modality. Cambridge London: Cambridge University Press. Pauliny, Eugène 1948. La phrase et l'énonciation. Recueil linguistique de Bratislava 1, 59–​66. Pauliny, Eugène 1967. La phrase et l'énonciation. In Vachek, Josef (ed) A Prague School Reader in Linguistics, Bloomington: Indiana University Press, 391–​397.

Inferential Evidentiality and Abduction

71

Peirce, Charles S., 1935. Collected Papers of Charles Sanders Peirce. Edited by Charles Hartsthorne /​Paul Weiss (1965/​1974), vol. 1–​6; Arthur Burks (1967/​1979), vol. 7–​8 edn. Cambridge (Mass.): The Belknap Press of Harvard University Press. Plungian, Vladimir A., 2001. The place of evidentiality within the universal grammatical space, Journal of Pragmatics 33/​3, 349–​357. Pollock, John L. 1987. Defeasible reasoning. Cognitive Linguistics 11/​4, 481–​518. Pólya, George 1949. Preliminary remarks on a logic of plausible inference. Dialectica 3/​9 –​10, 28–​35. Pólya, George 1989. Comment poser et résoudre un problème (trad. fr. de "How to solve it", Princeton, University Press, 1945). Sceaux: Éditions Jacques Gabay. Provôt, Agnès /​Desclés, Jean-​Pierre 2012. Existe-​t-​il un conditionnel médiatif en français? Faits de langue 40, 45–​52. Reiter, Raymond 1978. On reasoning by default, 1978, In Proceedings of Theoretical Issues in Natural Language Processing–2 (July 1978), 210–​218. Reiter, Raymond 1980. A logic for default reasoning. Artificial Intelligence 13/​1–​2, 81–​132. Reiter, Raymond 1987. Nonmonotonic reasoning. Annual Review of Computer Science 2/​1, 147–​186. Remnant, Peter 1948. An Evaluation of Charles Peirce's Concept of Retroduction, University of British Columbia. Rescher, Nicholas 1976. Plausible Reasoning. Assen: Van Gorcum. Rescher, Nicholas 2006. Presumption and the Practices of Tentative Cognition. New York: Cambridge University Press. Rigotti, Eddo 2009. Whether and how classical topics can be revived within contemporary argumentation theory. In Van Eemeren, Frans H. /​Garssen, Bart (eds) Pondering on Problems of Argumentation, Amsterdam: Springer, 157–​178. Rigotti, Eddo /​Greco, Sara 2019. Inference in Argumentation. A Topics-​ Based Approach to Argument Schemes. Cham: Springer. Rocci, Andrea 2006. Pragmatic inference and argumentation in intercultural communication. Intercultural Pragmatics 3/​4, 409–​442.

72

Dendale and Miecznikowski

Rocci, Andrea 2017. Modality in Argumentation. A Semantic Investigation of the Role of Modalities in the Structure of Arguments with an Application to Italian Modal Expressions. Dordrecht: Springer. Rossari, Corinne /​Cojocariu, Corina /​Ricci, Claudia /​Spiridon, Adriana 2009. The polysemy of devoir: A contrastive and diachronic analysis. In Rossari, Corinne /​Ricci, Claudia /​Spiridon, Adriana (eds) Grammaticalization and Pragmatics: Facts, Approaches, Theoretical Issues, Bingley: Emerald, 115–​136. Roulet, Eddy 1996. Une approche modulaire de l’organisation topicale d’un fragment d’entretien. Cahiers de linguistique française 18, 11–​32. Ruskan, Anna 2012. Evidential adjectives in Lithuanian academic discourse. Kalbotyra 64/​3, 103–​123. Squartini, Mario 2008. Lexical vs. grammatical evidentiality in French and Italian. Linguistics: An Interdisciplinary Journal of the Language Sciences 46/​5, 917–​947. Van Eemeren, Frans H. /​Grootendorst, Rob 1992. Argumentation, Communication, and Fallacies: A Pragma-​dialectical Perspective. Hillsdale, N.J.: Lawrence Erlbaum. Vet, Co 1988. Compte rendu critique de Wallace Chafe et Johanna Nichols, 1986. Canadian Journal of Linguistics 33, 65–​77. Vetters, Carl 2012. Modalité et évidentialité dans pouvoir et devoir: typologie et discussions. Langue française 173/​1, 31–​47. Walton, Douglas 1997. Appeal to Expert Opinion. University Park: Pennsylvania State Univ. Press. Walton, Douglas 2001. Abductive, presumptive and plausible arguments. Informal Logic 21/​2, 141–​169. Walton, Douglas /​Reed, Christopher /​Macagno, Fabrizio 2008. Argumentation Schemes, Cambridge: Cambridge University Press. Willett, Thomas 1988. A cross-​linguistic survey of the grammaticization of evidentiality. Studies in Language 12/​1, 51–​97.

Tabea Reiner

Further Motivating the Challengeability Criterion for Evidentiality1

Abstract At least since Faller (2002), challengeability has been used as a criterion to set apart epistemic modality from evidentiality: epistemic meanings can be challenged while (purely) evidential ones cannot (e.g. Murray 2017). However, the criterion is hardly established outside of Formal Semantics and has been criticized even within this approach. The present contribution argues that, in fact, challengeability receives compelling conceptual support from a tense-​aspect-​mood framework in the spirit of Klein (1994). After elaborating this approach, it is compared to Anderson’s (1986) and Nuyts’s (2017) works, which involve partially similar ideas. The discussion then turns to a potential fallacy: if temporal and aspectual as well as modal (epistemic) meanings are challengeable, are they necessarily lexical as opposed to grammatical entities according to Boye & Harder’s (2012) conception of the two spheres? I argue that, on closer inspection, it is not the temporal, aspectual or modal (epistemic) meaning as such that is challenged but the claim it serves to restrict –​which is still in contrast to evidentiality. Keywords: tense, aspect, mood, evidentiality, challengeability, at-​ issueness, lexicon vs. grammar.

1

I thank two anonymous reviewers for their comments and suggestions.

74

Tabea Reiner

1. Introduction: The TAM-​E Connection2 Research on tense (T), aspect (A), and mood/​modality (M) often comes under the heading of TAM (397 results in the Linguistic Bibliography Online) or TMA (39 results in the Linguistic Bibliography Online).3 Using the acronym makes sense, since the three categories and their respective subcategories are generally considered to be closely related. However, in recent years evidentiality (E) has also been admitted into the acronym, yielding TAM-​E or TAME (e.g. Molsing & Ibaños 2013, Ayoun et al. 2018; also cf. the references in Nuyts 2017: 57). Certainly, evidentiality is a well-​established notion in linguistics, usually said to involve the source of information for an utterance (Dendale & Tas­ mowski 2001: 339–​340). However, there is one respect in which T, A, and M are not only closely related but literally the same –​while E is different. The present paper aims to highlight this difference (section 2) and relates it to a well-​k nown criterion for identifying evidential meanings and their carriers: challengeability (section 3).

2. Why TAM is Fundamentally Different from E Asserting that one group of categories is different from some other category obviously hinges on how we define all of those categories precisely. Therefore, I will first introduce the framework used in the present paper, starting with its essentials from Klein’s (1994) seminal

2

3

The present contribution can be read as an independent motivation for Murray’s (2010, 2017) account of evidential meanings as non-​asserted and non-​at-​issue. For a contextualization of Murray’s approach as well as for several thought-​ provoking remarks, cf. Korotkova (2020). Since this paper has come to my atten­ tion only recently, I cannot do full justice to it but I will refer to it occasionally throughout the text. https://​bib​liog​raph​ies.bril​lonl​ine.com/​bro​wse/​lin​g uis​tic-​bibli​ogra​phy, accessed 24 September 2019

Motivating the Challengeability Criterion

75

work and proceeding with its further developments in later works by Klein (2008, 2009) as well as others including the present author. Only then it will be shown how the four categories relate to each other within the final model proposed here and in particular what distinguishes E from TAM. It should be noted that the framework is mostly concerned with types of meanings, less with the types of expressions that have these meanings. So identifying the latter must remain a minor topic until section 3. 2.1. The Framework Klein (1994). In his 1994 monograph Time in Language, Wolfgang Klein reinterprets Reichenbach’s (1947: 287–​298) reference time (R) as topic time (TT), i.e. the time for which a claim is made (p. 4). For example, consider a witness who utters (1). (1) The light was on. (Klein 1994: 2)

This testimony makes a claim exclusively about a certain period of time in the past; it is immaterial whether the light is still on at the time of utterance (pp. 2–​4). Against this background, Klein (1994: 6) suggests the following definitions : •

tense

•

=​relation between topic time (TT) and time of utterance (TU) aspect =​relation between topic time (TT) and time of situation4 (TSit)

I consider this minimal grid the essence of Klein (1994). Defending it in full against other approaches, especially with respect to non-​assertions,

4 By situation, Klein (1994: 3) means whichever event or state is denoted by the verb together with its arguments. Apart from this specification, the terms “time of situation”, “time of utterance” and “topic time” are to be understood holistically. In particular, Klein does not develop a theory of topicality.

76

Tabea Reiner

would go far beyond the scope of the present paper.5 However, I would like to indicate that the definitions, besides accounting for the standard cases, easily capture some seemingly deviant phenomena. Among these are certain uses of the simple past (ʻTT before TUʼ) with stative or copular verbs: TT before TU, TSit may overlap TU (2) I liked Madrid! (3) His name was Charles. (4) She was here and she still is.

Since tense locates TT (rather than TSit) relative to TU, the intended interpretations of the above examples fall out naturally: the speaker claims the occurrence of the respective situation for a time before TU, leaving underspecified whether the situation continues to hold at TU. If, by contrast, tense was said to locate TSit, then the above examples would have as their literal meaning that the liking/​name carrying/​being on-​site is over at TU –​which in the case of (4) would result in an outright contradiction (also cf. Iatridou 2000: 248 for a similar example). Likewise, aspect locates TT –​albeit not relative to TU but rather relative to TSit. Crucially, neither tense nor aspect specifies the relation between TU and TSit directly, as traditional grammar would have it. So, the Kleinian approach involves a certain amount of underspecification. In spite of this last point, however, both tense and aspect do imply a restriction: in locating TT, they specify the time for which a claim is made, i.e. via them the claim is restricted or anchored to that time. Note that underspecification and restriction do not conflict here since a restricted claim leaves open the possibility that any of its less restricted versions is true as well. It is this restricting effect of tense and aspect that will become central later in the paper. Further Developments. Thus far, only tense and aspect have been addressed –​so what about mood/​modality? In fact, it can be defined in an analogous way, provided two additional terms, i.e. TW and WSp:

5

For example, a comparison with Leiss (2012) –​an approach with related but different motivations and different outcomes –​would be an appealing task.

Motivating the Challengeability Criterion

77

mood =​relation between TW (topic world, i.e. the world for which a claim is made) and WSp (actual world according to the speaker)

It should be noted that for present purposes nothing hinges on the choice of terms, i.e. mood vs. modality. In particular, the decision for mood made here is not supposed to entail a certain way of marking the relation or to reflect specific theoretical persuasions. That is, I include the meanings of conditional inflections as well as the meanings of epistemic modal verbs (i.e. the relevant part of these meanings) and I do not take a definite stance on whether or where to locate the corresponding constituents in a syntactic model. However, the notion of mood proposed above excludes the meanings of dynamic and deontic modal verbs to the extent that these meanings do not relate to possible worlds (as frames of reference for truth). The basic idea of topic worlds was, to my knowledge, first pursued in Formal Semantics (Iatridou 2000: 247, Gergel 2001: 138) and afterwards in Klein’s later works (2008: 289, 2009: 340–​1), where the topic world is just one ingredient of a broader “topic situation/​component”. The latter cannot be taken into account in full within the present paper but note that such a broader perspective relates in interesting ways to Ritter & Wiltschko’s (2009) take on the syntax of finiteness. What I would like to do here instead is to combine the semantic notions introduced so far (TT, TU, TSit, TW, WSp) into an integrated framework, thereby refining a proposal in Reiner (2019: 305). This is done in Figure 1, followed by further explanations. It should be noted that this figure provides the definitional foundation for the present paper in that it involves a graphical rendering of all relevant definitions given so far as well as additional definitions to be addressed below. Furthermore, it can be read as an inheritance hierarchy from “comment” downwards. This framework, as well as its predecessor in Reiner (2019), was originally designed for providing a semantic notion of (non-​)finiteness by exploiting the central Kleinian insight: to claim or not to claim, that is the question. Due to this origin of the framework, Figure 1 features both a non-​finiteness branch on the left and a finiteness branch on the right. The former involves two new terms, i.e. TX and WX, denoting some specific time/​world, respectively. More important for purposes of the

78

Tabea Reiner

present paper however, is the right hand side of Figure 1. It comprises the three terms tense –​ aspect –​ mood as defined in prose above and it highlights how these definitions relate to utterance meaning as a whole. Given a topic-​comment division, the comment may constitute a claim about the topic. Claims, however, are always or at least typically claims for a certain domain: even a general statement like Two plus two makes

Figure 1:  Framework of semantic TAM-​definitions

four does not come without a domain but, instead, implies that the domain is a very broad one, i.e. ʻall timesʼ (Klein 1994: 6) and ʻall consistent worlds with the Peano axiomsʼ.6 More obviously, specific statements like Partee’s (1973: 602) famous example I didn’t turn off the stove or the examples referred to above are confined to certain times

6

I consider it important to distinguish between consistent and inconsistent worlds, unlike practitioners of Possible Worlds Semantics (Reiner 2018).

Motivating the Challengeability Criterion

79

and worlds. Incidentally, claims can even be confined to worlds that the speaker considers unreal like in I would travel the world [if I was rich]. It is this restricting effect that defines tense, aspect, and mood in Figure 1 above. Retranslated into prose, the definitions read like this: • • •

If the claim is restricted to a certain world TW and TW is identified by its relation to WSp then the latter relation is called mood. If the claim is restricted to a certain time TT and TT is identified by its relation to TU then the latter relation is called tense. If the claim is restricted to a certain time TT and TT is identified by its relation to TSit then the latter relation is called aspect.

Needless to say that both sorts of restrictions –​to TW (mood) as well as to TT (tense, aspect) –​may but need not apply at the same time. Nor is it required that TT’s two relations are always specified in tandem. To sum up, mood identifies TW, tense and aspect specify TT, together (in any given combination) they specify the domain to which the claim is restricted. This is the role of TAM that will become crucial in the next subsection. 2.2. The Difference between TAM and E according to the Framework If tense, aspect, and mood (TAM) crucially help restrict the claim –​ what is the role of evidentiality (E) then? At first sight, it seems to serve exactly the same function. This impression, however, is only partially correct. Consider the delimitation of evidentiality in a recent handbook, here picking out only the semantic components. According to Aikhen­ vald (2018: 1–​5), evidentiality involves the expression of... ...ways of acquiring information ...and nothing else =​information source

Also according to Aikhenvald (2018: 5–​6), evidentiality does not inher­ ently involve the expression of...

80

Tabea Reiner ...justification7 ...truth ...belief ...level of certainty ...level of reliability

From this perspective, which I adopt here, evidentiality is really only about the speaker specifying how (s)he came to have the respective piece of information, e.g., through watching, hearing, or inferring.8 Crucially, acquiring information comes logically before recruiting it as a basis for making some claim. Thus, evidentiality does not operate on claims but rather on potential bases for claims. These conceptual distinctions become especially apparent in examples where the source is not judged reliable by the speaker, such as (5) from Cuzco Quechua.9 (5)  Cuzco Quechua (Faller 2002: 191) Pay-​kuna-​s (s)he-​PL-​si

ñoqa-​man-​qa I-​illa-​top

saqiy-​wa-​n, leave-​1o-​3

mana-​má not-​surp

7

8

9

riki right

qulqi-​ta money-​acc

muntu-​ntin-​pi lot-​i ncl-​loc

riku-​sqa-​yki see-​pp-​2

ni not

un one

I assume that Aikhenvald’s notion of justification is different from Anderson’s (1986) one, to be introduced below in n. 17. Aikhenvald seems to mean ʻvindicationʼ (Aikhenvald 2018: 5). I stick here with a traditional perspective, from which evidentiality is about the actual speaker’s information source (Anderson 1986: 277). This perspective has been widened by, amongst others, Nuyts (2001: 72–​78) but will not play a role in the discussion of Nuyts (2017) to follow later in this paper (cf. Nuyts 2017: 58, fn. 2). For more on reliability as a pragmatic (i.e. cancellable) link between evidential and epistemic meanings, cf. Cornillie (2009) as well as Wiemer (2018: 99–​102). For the complexities of Quechuan evidentiality and related categories cf. How­ ard (2012), Grzech (2016), Hintz & Hintz (2017) as well as Nuckolls (2018).

Motivating the Challengeability Criterion

sol-​ta Sol-​acc

centavo-​ta-​pis cent-​acc-​a dd

81

saqi-​sha-​wa-​n-​chu leave-​prog-​1o-​3-​n eg

ʻThey left me a lot of money, but, as you have seen, they didn’t leave me one sol, not one cent.ʼ (ev: It is said/​they said that they left me a lot of money.) (spontaneous)10

So the fundamental difference between TAM and E is this: while T, A, M are, by definition, assistants in claim-​shaping, the definition of E does not draw on claims directly. This difference can be modeled by integrating the definition of evidentiality into Figure 1 (also cf. Déchaine et al. 2017, Murray 2017), which results in Figure 2.

Figure 2:  TAM vs. E

10 Abbreviations in order of occurrence (Faller 2002: xii–​xiv): pl =​plural, illa =​illative, top =​topic, acc =​accusative case, incl =​inclusive, loc =​locative case, 1o =​first person object recipient, 3 =​third person, surp =​surprise, pp =​past participle, 2 =​second person, add =​additive, prog =​progressive, neg =​negative, ev =​evidential value.

82

Tabea Reiner

Note that evidentiality is inserted before the finiteness-​divide, so it is predicted to be compatible in principle with semantically non-​ finite contexts. As to the subcategories of evidentiality, I remain mostly agnostic. Aikhenvald (2018: 12) lists the following ones: visual, non-​ visual sensory, inference (from visible or tangible result), assumption (not from visible or tangible result but, e.g. from reasoning), reported, and quotative. Others, including the present author, might exclude quotative for being too specific: in contrast to reported it does not only specify the way of acquiring information (cf. the delimitation of evidentiality above) but presupposes a maximally specific origin of information, i.e. the quotee (Boye 2012: 32, more generally: Murray 2017: 10–​11).11 So the lower boundary of the evidential subcategories may be subject to debate. The upper boundary, however, appears to be consensual: if evidentiality is, as just emphasized, about ways of acquiring information then very broad meanings such as ʻaccording to some information that the present speaker has acquiredʼ are too general. For carriers of these meanings –​e.g., verbs like English turn out, Dutch blijken, French s’avérer, and Spanish resultar –​ Dendale (2019) reserves the term discovery markers. As a whole, the subcategories of evidentiality and their boundaries are relevant for getting a lucid idea of the category but will only play a minor role in the rest of this paper, except for the discussion of Nuyts (2017) to follow in section 3.3. To sum up section 2, evidentiality (E) as well as tense, aspect, and mood (TAM) were defined from a semantic point of view whereby the former turned out to be fundamentally different from the latter. The next section addresses the question of how the definitions may be exploited for identifying markers of E vs. TAM in text. It will be this section that relates the special status of evidential meanings to their non-​challengeability. Although the section is about markers, i.e. about signifiants, the method will be purely semantic as the definitions are.

11 I thank an anonymous attendee of the workshop Evidentials versus Non-​ Evidentials (16th IPrA meeting 2019, Hong Kong) for making me aware of this distinction.

Motivating the Challengeability Criterion

83

3. Relating the Special Status of E to the Challengeability Criterion 3.1. From Claims to Challengeability to Determining the Absence of Either Claims can be challenged; in particular, a claim for a certain TW and/​ or TT can be challenged. That is: given a statement in discourse, an interlocutor may felicitously call into question the statement’s validity for the world and/​or time it is confined to. For instance, the world might be WSp (actual world according to the speaker, in the simplest case being identical in principle to the actual world according to the interlocutor) and the time might be, to revisit example (2), the time span of a conference a few years ago: (6)

A: I liked Madrid! B: You didn’t, remember, you kept complaining about the heat.

Here B’s reaction could be paraphrased as ʻFor what both of us usually consider reality and for the time we are talking about, your statement does not holdʼ. It should be noted that in this example as well as in general, the challenged part is not mood, tense, or aspect as such –​ you cannot challenge a relation –​but the claim for a certain world/​time, which is identified by the relations. How does this differ from evidentiality, as defined above? With evidentiality, there is no claim, which could be challenged, cf. Figure 2. At most, the speaker’s source provides pieces of information yet to become vindications for future claims. So, given a statement in discourse, an interlocutor may felicitously challenge the statement –​even based on the perceived reliability of the information source –​but not the mere way of acquiring information. This property of evidential meanings identifies them in text and with them it identifies their carriers. Put differently: non-​challengeability is a necessary (though not

84

Tabea Reiner

sufficient) criterion for identifying evidential markers.12 Here is an oft-​ cited example from Cuzco Quechua: (7)

Cuzco Quechua (Faller 2006: 11, presented below essentially following Murray 2017: 12) A:  Ines-​qa Inés-​top

qaynunchay ñaña-​n-​ta-​s yesterday sister-​3-​acc-​r ep

watuku-​sqa. visit-​pst2 p(roposition) =​‘Inés visited her sister yesterday.’ ev =​speaker was told that p B:  Mana-​n not-​bpg

chiqaq-​chu. Manta-​n-​ta-​lla-​n true-​n eg mother-​3-​acc-​lim-​bpg

watuku-​rqa-​n. visit-​pst1–​3 ‘That’s not true. She only visited her mother.’ B’:  Mana-​n not-​bpg

chiqaq-​chu. true-​n eg

#Mana-​n chay-​ta willa-​rqa-​sunki-​chu. not-​bpg this-​acc tell-​pst1-​3s2o-​n eg ‘That’s not true. You were not told this.’ (elicited)13

12 More schematically: (meaning evidential → NOT(meaning challengeable)) ↔ (NOT(meaning evidential) ← (meaning challengeable)) Crucially, I do not take a stance on what follows from NOT(meaning evidential). In particular I do not hold that all epistemic meanings are necessarily challengeable in all circumstances. 13 Data type according to Faller (2006: 21). Abbreviations in order of occurrence (ibid.): top =​topic, 3 =​third person, acc =​accusative, rep =​reportative, pst2 =​ past tense 2, bpg =​best possible grounds, neg =​negative, lim =​limitative, pst1 =​ past tense 1, 3s2o =​third person subject second person object.

85

Motivating the Challengeability Criterion

Next is a parallel example from Cheyenne : (8) Cheyenne (Murray 2017: 13) A: Méave'ho'eno é-​héstȧhe-​sėstse Lame Deer 3-​be.from-​r pt.3sg Mókéé'e is from Lame Deer, I hear.ʼ

Mókéé'e. Mókéé'e

B:  É-​sáa-​hetómėstovė-​hane-​ø. 3-​not-​be.true-​n eginan-​w tn É-​sáa-​héstȧhé-​he-​ø Méave'ho'eno. 3-​not-​be.from-​n egan-​w tn Lame Deer ʻThat’s not true. She’s not from Lame Deer.ʼ B’:  #É-​sáa-​hetómėstovė-​hane-​ø. 3-​not-​be.true-​n eginan-​w tn Hovánee’  eé-​sáa-​nė-​hé-​he-​ø. nobody   3-​not-​that-​say-​n eginan-​w tn ʻThat’s not true. Nobody said that.ʼ (presumably elicited)14

Provided that the forms glossed as reportative in (7) and (8) do, in fact, convey evidential meaning in the sense specified above (cf. Fig­ ure 2), their discourse behavior does not come as a surprise: since evidential meanings do not involve claims, there is nothing that could be challenged –​so the alternative reaction by B is bound to be inappropriate, as predicted above. Against this background, it seems natural that challengeability has been used in the literature as a test to identify evidential markers. Even more so, the criterion falls out automatically from Figure 2. Thus, 14 The example is cited in Murray (2017: 13) from Murray (2010: 51) and in the original work (p. 10) she points out: “[m]‌ost of the examples from my fieldwork come from elicitation sessions or (modified) texts, but some are examples that I observed or overheard, or heard in a Cheyenne language class”. Abbreviations in order of occurrence (Murray 2017:xv): 3 =​third person, rpt =​reportative, sg =​singular, neg =​agreement that appears with negation and the inferential evidential, inan =​inanimate, wtn =​direct (witness) evidential, an =​animate.

86

Tabea Reiner

anyone who accepts the basic distinctions drawn there must as well accept [-​]‌challengeability as a necessary (not sufficient!) criterion for evidential meanings. 3.2. Doubts about the Criterion of Challengeability In spite of what has been said above, the challengeability criterion has so far received little support beyond Formal Semantics. For example, Boye’s (2018) overview article on the notion and the term of evidentiality does not even mention it, which is connected with a certain reluctance to draw an absolute distinction between evidentiality and modality at all.15 In fact, even within Formal Semantics, there are reasons to suspect that the criterion of challengeability is not valid. More precisely, evidentiality as defined in the present paper seems to be one kind of non-​at-​issue meaning (cf., e.g. Gutz­ mann & Turgay 2019) and we should expect all of these meanings to be non-​challengeable. The problem here is not that the criterion applies too broadly: as long as challengeability is viewed as merely a necessary –​not sufficient –​criterion for evidentiality, the criterion is expected to apply to other categories as well (cf. Korotkova 2020 on self-​attributions more generally). The real problem is that some non-​ at-​issue meanings may have turned out to be challengeable, at least under certain conditions. For example, Müller (2019a: 315) presents the following dialogue: (9) A: B:

Mr. Peanutbutter, a talking horse, is a successful actor. Hey, wait a minute, Mr. Peanutbutter is not a talking horse, he’s a Labrador.

In A’s utterance, the proposition ʻMr. Peanutbutter is a talking horseʼ (i.e. the proposition introduced by the parenthetical) is not 15 Boye (2018: 264), referring back to Boye (2012), does draw a distinction between “evidentiality and epistemic modality as cross-​linguistic descriptive categories” but adds a “superordinate category of epistemicity”. The latter is incompatible with the approach defended in the present contribution.

Motivating the Challengeability Criterion

87

at-​issue, let alone part of the claim –​it is merely a presupposition, cf. its survival under negation: Mr. Peanutbutter, a talking horse, is not a successful actor. Nonetheless, it seems that this meaning can be felicitously challenged by B in (9) above. The crucial formula here is Hey, wait a minute, i.e. one of the many variants of Shanon’s (1976) test environment for identifying what he calls pragmatic presuppositions (also cf. Müller 2019a: 314). Shanon distinguishes these from what he calls semantic presuppositions. Though, the latter appear to be mere entailments (p. 249). So it is for presuppositions (and their triggers) in general that Shanon’s test was adopted by the community (e.g. von Fintel 2004: 317, Matthewson 2006). When Tonhauser et al. (2013: 81) review its use critically then it is only because in an elicitation setting, the test may produce false negative as well as false positive results, due to the intricacies of elicitation as such. To sum up for the moment, there is an established test for identifying a certain kind of non-​at-​issue meanings (i.e. presupposition) and this test is based on challenging these meanings. How can this be, how can we challenge meanings that are not even at-​issue? I hold that the mystery is solved if we realize that the respective formula –​in (9) above this was Hey, wait a minute –​conveys ʻI want to talk about something different for a momentʼ. Thus, it changes what is at issue. This can be shown by a repeated application of the test: (10) A: B: B':

Hey, wait a minute, Mr. Peanutbutter is not a talking horse, he’s a Labrador. Oh, I did not know that he’s a Labrador. #Hey, wait a minute, I did not know that he’s a Labrador.

This property of the formula is certainly not a problem for Shanon’s diagnostic or its descendants; rather, it shows that what seemed to be a challenge of non-​at-​issue meanings is, in fact, a challenge of former non-​at-​issue meanings after they have been made at-​issue meanings.16 16 I am assuming here a proposal-​based approach as opposed to a question-​based approach (Gutzmann & Turgay 2019: 13–​18) insofar as at-​issueness is tied to a certain class of expressions (e.g. Hey, wait a minute). On a related note, I suspect Bavarian fei (Hinterwimmer 2019) to have a similar effect as Hey, wait a minute in that it selects a non-​at-​issue meaning and turns it into an at-​issue meaning.

88

Tabea Reiner

In conclusion, the notion that non-​at-​issue meanings are not challengeable is rescued. This pertains to both presuppositions as well as evidential meanings: if also the latter turn out to be “challengeable” by Shanon’s test then this behavior is readily captured by the insight that also these meanings have been dragged to the negotiable surface of discourse by one of the formulas. Put differently, there is no contradiction between indirect challengeability and the definition of evidentiality as depicted in Figure 2. This perspective avoids the far-​reaching conclu­ sion that, in light of examples like (9) above, challengeability is essentially a void criterion for evidentiality, as drawn by Müller (2019b). Another challenge for challengeability will be discussed in the subsection after next (cf. example (14) there). Before that, however, I intend to indicate the roots of the challengeability criterion (Anderson 1986) and give an example of a modern approach that does without it (Nuyts 2017). In the latter case in particular, this amounts to a full scale comparison with the approach suggested in the present contribution. 3.3. Comparison of the Present Approach to Anderson (1986) and Nuyts (2017) Anderson (1986). The definitions and relations depicted in Figure 2, implying the criterion of challengeability (see above), have an important forerunner in Anderson’s (1986) work on evidentiality. Proceeding from a set of expressions that mark information source,17 he recognizes four criteria (a–​d) for “archetypal evidentials”, including b): ‌

Evidentials are not themselves the main predication of the clause, but are rather a specification added to a factual claim about something else. (Anderson 1986: 274, emphasis original; the exposition as a whole extends from p. 273 to p. 276, cf. also p. 311)

17 Information source is a modern rendering of Anderson’s (1986) wording “kind of evidence a person has for making a factual claim” (p. 273) or “kind of justification for a factual claim which is available to the person making that claim” (p. 274, cf. also Boye 2018: 269). Note that the claim as such is left untouched.

Motivating the Challengeability Criterion

89

Thus, already Anderson states that evidential meanings reside outside of the claim proper and are non-​at-​issue, hence18 not directly chal­ lengeable. The characteristic that these meanings are non-​ at-​ issue, incidentally, makes their markers grammatical markers as opposed to lexical ones from Boye & Harder’s (2012) perspective (cf. also Boye 2018: 269–​270). This is a point to which I will return later. Coming back to Anderson’s (1986) criterion b) cited above, what does the pres­ ent approach have to add to this criterion? Both Anderson’s criterion and the present approach permit the conclusion that evidential meanings cannot be challenged. However, they arrive at this conclusion from opposite points of departure: while Anderson (1986: 273) deliberately avoids “proceed[ing] from a priori logical ideas” and his criteria are, in fact, empirical a posteriori findings, the present approach derives the criterion of challengeability from a comprehensive framework of TAM-​vs. E-​semantics. Both lines of research taken together constitute converging evidence for the criterion of challengeability. Nuyts (2017). The present contribution as a whole bears an obvious resemblance to Nuyts (2017) and previous work by the same author in that both approaches criticize the use of the acronym TAM-E/​ TAME. Additionally, there is a similarity with respect to the treatment of deontic modality (see below). However, on the whole, Nuyts works in a markedly different theoretical framework and arrives at far more radical conclusions than I do. In order to assess the commonalities and differences of the two approaches in more detail, the following paragraphs provide an outline of Nuyts (2017) with commentary. Nuyts (2017: 61) starts from the following hierarchy of qualifica­ tional categories:

18 Note that even this relation may be disputed; cf. Korotkova (2020), exx. (11) and (12). For the rest of the present paper, I take it that non-​challengeability follows from non-​at-​issueness, which is equivalent to saying that at-​issueness follows from challengeability. Moreover, I assume that the reverse also holds; so in sum, challengeability and at-​issueness are treated as equivalent here. In addition, the trivial assumption is made that claims are challengeable.

90

Tabea Reiner

> evidentiality   > epistemic modality   > deontic modality > time     > quantitative aspect/​dynamic modality      > phasal aspect > (parts of the) state of affairs

The left angle bracket means ʻscopes overʼ (with a qualification that need not concern us here, cf. Nuyts 2017: 68, fn. 6) and I take the top most bracket to mean that also the top layer is scoped over by something, i.e. by discourse categories having the qualificational hierarchy as their input. This hierarchy is slightly modified in the course of Nuyts’ argument and I will present the final version below. In any case, the hierarchy is a conceptual one; thus, in its linguistic interpretation, it is decidedly not restricted to cover only those meanings that are conveyed by grammatical means (Nuyts 2017: 60–​61). Zooming in on evidentiality, there is likewise no restriction to non-​at-​issue meanings (Nuyts 2017: 67, ex. 5a). Comparing this theoretical framework to the one sketched in the present contribution (cf. Figure 2 above), there appears to be a partial similarity with respect to the graphic arrangement of categories: evidential above modal above temporal. However, apart from the fact that such an arrangement is quite commonplace (cf., e.g. work in cartographic syntax like Rizzi & Cinque 2016: 149), the motivations as well as the consequences of the two models differ considerably. As to the motivations, Figure 2 is not about scope but about specification (from “comment” downwards). As to the consequences, recall the implications of Figure 2 in tandem with Anderson’s criterion b): the former locates evidential meanings outside of claim formation, which predicts that they are not challengeable (and non-​at-​issue); the latter states that evidential meanings are non-​at-​issue (which entails that they are not challengeable). Nuyts’ framework, however, eschews restrictions of this sort. So with respect to evidentiality, there is a marked difference in restrictiveness between the two models compared here (Nuyts’ initial and Nuyts’ final model are the same in this respect).

Motivating the Challengeability Criterion

91

The difference may extend to the lexical vs. grammatical expression of evidentiality (recall that Nuyts includes both); however, any discussion of this issue requires some background premises that I cannot introduce until the section after next (“The Inevitable Question: Are we talking about Lexical or Grammatical Realizations of TAM and E?”). So suffice it to say that Nuyts (2017) opts for an all-​encompassing per­­ spective while the present framework will be shown to exclude certain phenomena. Even so, the points of departure are quite similar: Nuyts starts from conceptual considerations, I start from semantic considerations, in none of the approaches do combinatorial considerations play an initial role. It is shortly afterwards that very different routes are taken. As an interim conclusion, the two frameworks do diverge to a substantial degree so that the specifics hardly seem to be comparable. However, some of Nuyts’ (2017) core points –​motivating his subse­ quent modification of the hierarchy –​concern the category of interest here, i.e. evidentiality (with respect to its potential subcategories). So I will summarize and discuss these points in what follows. His main hypothesis is that inferential evidentiality patterns with deontic and epistemic modality whereas experienced and hearsay do not.19 This is the reason why he criticizes the use of the acronym TAM-​E/​TAME: if his hypothesis is correct, then only one of evidentiality’s traditional subcategories is truly on a par with tense/​time, aspect/​ aspectuality, and mood/​modality; i.e. inferential evidentiality. So, how does Nuyts support his hypothesis? He gives essentially two reasons: 1. 2.

Inferential evidentiality is “scalar” (as are deontic and epistemic modality) while experienced and hearsay are not (Nuyts 2017: 69–​70). Inferential evidentiality is “speaker-​oriented” (as are deontic and epistemic modality) while experienced and hearsay are not (Nuyts 2017: 70–​71).

19 Experienced corresponds to visual +​non-​v isual sensory from Aikhenvald’s (2018: 12) list, inferential corresponds to inference +​assumption from that list, and hearsay corresponds to reported.

92

Tabea Reiner

I do not have to say much about reason number 2: within the framework of the qualificational hierarchy, this reason is indeed a justification for splitting the evidential category, since the hierarchy is also about increasing involvement and effort on the side of the speaker (Nuyts 2017: 62, 66) –​whereas within Figure 2 above, these dimensions do not play a role; they are, as it were, “abstracted away”. However, reason number 1 is worth closer examination. What Nuyts (2017: 69) says is, more precisely, that inferential evidentiality is associated with a scale of reliability (also cf. Wiemer 2018), while expe­ rienced and hearsay simply specify the kind of source without such an association. I have three comments here, one global and two local. The global comment is that, strictly speaking, Nuyts (2017) does not compare the actual categories: as becomes very clear from his own explanations (cf., e.g. his ­figure 7 on p. 72), it is not inferential evidentiality as such that is scalar but it is the associated category of reliability, which experienced and hearsay appear to lack. To put it bluntly: just as we do not see things to different degrees, we do not infer them to different degrees –​degree of certainty is but a judgment on the result of inferencing. This is why I see no reason to treat inferential evidentiality as such differently from the other subcategories. My two more local comments both concern the fact that reason number 1 (now to be read as “…is associated with scalarity…”) is an empirical one (cf. p. 72, fn. 8). First, this means that the quest for counterexamples is open. A potential counterexample is provided by inferentials and conjectures20 in Matses (Panoan, Amazonian Peru and Brazil), which appear to be neutral with respect to reliability (Fleck 2007: 598, 602). Second, Nuyts’ category (purely-​)experienced would be ultimately justified –​and my category (purely-​)evidential securely filled –​by a type of example that is notoriously hard to come by: visual evidentials that are used even after the resolution of an optical illusion (recall that in the case of optical illusions perceptions are not altered by knowledge). In mock English, such an example could look like (11). (11) Line A is-​v isual longer than line B, but in fact, line A is as long as line B.

20 It is not clear to me whether conjectural meanings are included in Nuyts’ cate­ gory inferential, in any case he includes “logical reasoning […] from general background knowledge and assumptions” (Nuyts 2017: 67).

93

Motivating the Challengeability Criterion

Unfortunately, the only potential example that I am aware of, i.e. Cuzco Quechuan -​mi, does not turn out to behave in this way (Faller 2002: 154). This fact does not spoil either approach but I think it is useful to picture ideal examples in order to intensify looking for their matches in reality. On the whole, Nuyts’ (2017) treatment of experienced and hear­ say as different from inferential evidentiality reveals what is at the heart of our shared unwillingness to use the acronym TAM-E/​TAME: we delete the “real” evidential categories, i.e. those that are argued to be exclusively about the kind of information source, from the system of temporal-aspectual-modal meanings. The difference is that, for Nuyts, inferential evidentiality does not belong to the real evidential meanings whereas I could not follow his reasoning to this effect. Accordingly, his modified version of the qualificational hierarchy looks like this (Nuyts 2017: 73, emphasis mine): > inferential evidentiality > epistemic modality > deontic modality > time > quantitative aspect/​dynamic modality   > phasal aspect     > (parts of the) state of affairs

[attitudinal] [situating] [detailing]

Taking a second look at this hierarchy, yet another partial overlap with the present contribution can be seen in the treatment of deontic modality: as Nuyts (2017: 72) emphasizes, deontic modality is about “the moral status of a state of affairs, no matter whether hypothetical or real”. This characteristic sets off deontic modality from categories having to do with reliability or reality, i.e. from inferential evidentiality and epistemic modality within Nuyts’ final qualificational hierarchy as well as from mood within my own framework. However, the difference between the models is that deontic and also dynamic modality are still awarded a place in Nuyts’ hierarchy, whereas they are dispatched to some other domain of description by the logic of Figure 2 above. Rounding off my discussion of Nuyts (2017), his conclusions are far more radical than mine in that he does not even acknowledge evidentiality as a coherent category. As argued above, one of his two

94

Tabea Reiner

reasons for this decision does not carry over to the present framework and the other one actually concerns another category, i.e. reliability. To conclude the section on comparisons, the first one provided converging evidence for the criterion of challengeability; the second one provided converging evidence against the acronym TAM-E/​TAME (although even within the particular framework I could accept only one out of two reasons given for this). 3.4. Demonstrating the Difference between TAM and E with respect to Challengeability The formal-​semantics literature on evidentiality is full of dialogic examples where the truth value of an utterance can be challenged while trying to challenge the evidential meaning results in infelicity (cf., among others, Gutzmann & Turgay 2019: 3 or examples (7) and (8) given above from Faller 2006 and Murray 2017, respectively). What I intend to do in the present section, however, is slightly different. I want to show that not only truth values (which align with modal meanings in the prototypical case) can be challenged but that also temporal and aspectual meanings are separately challengeable, hence differ from evidential meanings in this respect –​as predicted by Figure 2. Since felicity judgments require native speaker intuitions, I will use the only language of which I have native command, i.e. German. German is not among those languages that are famous for possessing a full-​fledged system of evidentiality markers, let alone a grammatical system of such markers (Faller 2006: 4). However, there is one modal verb that regularly surfaces as a candidate-​evidential in the literature, i.e. (candidate-​)reportative soll-​ (e.g. Palmer 2001: 9; Faller 2006, 2017; Schenner 2008, 2010; Murray 2017: 14–​15). Other candidates include quotative woll-​(ibid.) as well as the two conditional paradigms in their quotative function. Even semi-​modals and certain readings of the future construction have been examined for evidential meaning (Diewald & Smirnova 2010). I will concentrate on (candidate-​)reportative soll-​in the rest of the current subsection, excluding from consideration the quotatives as well as the semi-​modals and the future readings. The reason has to do with the lower and upper

95

Motivating the Challengeability Criterion

boundaries of evidentiality briefly addressed above in the context of evidential subcategories (shortly after Figure 2): while quotatives run the risk of being too specific, semi-​modals and the future readings run the risk of being too general (also cf. Ramat 2012). For the present pur­ poses, I will assume that a) (candidate-​)reportative soll-​is not just a candidate but indeed conveys (reportative) evidential meaning in the sense of Figure 2 and b) certain other elements convey temporal, aspec­ tual, or modal meaning in the sense of Figure 2. As a point of departure, consider the following (modified) example from the literature: (12)

German (adapted from Faller 2006: 12)21 A: Die Gärtnerin soll die the female.gardener rep the

Juwelen jewels

gestohlen hab-​en. steal.ptcp aux-​i nf ʻThe gardener is said to have stolen the jewels.ʼ B: #Nein, das hast du nicht gehört/​gelesen. (E) no that aux you not hear/​read.PTCP ʻNo, you haven’t heard/​read that.ʼ

(12) is exactly parallel to the Cuzco Quechua and Cheyenne examples: it shows that the way of acquiring information cannot felicitously be challenged.22 This is in large agreement with the literature on German (e.g. Faller 2006: 12, Murray 2017: 14–​15, from a slightly different point of view: Vanderbiesen 2018). However, it is important to note that B’s reac­ tion aims at the way of acquiring information exclusively. So it is really only evidential meaning that is put to the challengeability test here (also cf. Déchaine et al. 2017: 39). 21 Glossing in examples that are not directly borrowed from the literature follows the Leipzig Glossing Rules ; additional abbreviations: rep =​reportative, prt =​particle. 22 As with any test, there is the danger that the results are an artefact of other fac­ tors, which have not been sufficiently considered. Yet, in this case, I do not see any candidate for such a factor.

96

Tabea Reiner

I suspect that some applications of the challengeability test in the literature suffer from lack of precision in this regard and hence do not produce reliable results. A case in point might be Matthewson’s (2012: 104, ex. 40) argument that St’át’imcets lákw7a is not a sensory non-​visual evidential but a modal: in her dialogic example, the initial challenge actually targets the truth value of the original utterance and only then, in two separate orthographic sentences with no elements (pronominal or otherwise)23 discernibly referring back to the original utterance, is the way of acquiring the information discussed. Incidentally, exactly this example was heavily criticized by Aikhenvald (2018: 7–​8), although for different methodological reasons: Aikhenvald seems to equate elicitation with translation tasks and condemns them altogether. Coming back to example (12), it should be noted that everything said above refers to the relation between soll-​’s evidential meaning and the embedded proposition, i.e. the embedded potential claim: it is this claim for which soll-​specifies the type of information source. On this layer, the evidential meaning is outside of the claim, hence not challengeable. However, on a second layer, soll-​’s meaning does not convey the type of information source for any given claim but instead constitutes a claim of its own: ʻI claim that it is from hearsay [that I can claim…]ʼ (cf. also the finite morphology). The crucial thing to note is that this is different in scope and claim-​relatedness from evidentiality as depicted in Figure 2: it is, as it were, second-​order evidentiality. Also Korotkova’s (2020) example (16) from Dutch, Het SCHIJNT te regenen “It seems to rain” belongs here. Distinguishing between first-​ and second-​order evidentiality will be at the heart of distinguishing between “lexical” and “grammatical” expressions of evidentiality, cf. the next section. As an interim summary, challenges that really only target first-​ order evidential meanings are infelicitous. Let’s compare this situation to challenges that target the temporal, aspectual, or (for the sake of

23 I have to guess two of Matthewson’s glosses since these are not resolved in the paper: sbjn → new subject, ynq → yes/​no-​question.

97

Motivating the Challengeability Criterion

completeness) modal meaning of the original utterance. To keep things simple, I will use a slightly altered version of (12): (13) German (constructed) A: Die Gärtnerin hat die the female.gardener aux the

Juwelen jewels

gestohlen. steal.ptcp B1:

Nein, sie hat sie no she aux them ʻNo, she hasn’t stolen them.ʼ

B2: Nein, sie wird sie no she will them stehl-​en. steal-​i nf ʻNo, she will steal them (at a later time).ʼ B3: Nein, sie ist no she is ʻNo, she is at it right now.ʼ

gerade right.now

nicht not

gestohlen . (M) steal.ptcp

erst prt

noch prt (T)

dabei. at it

(A)

B’s first reaction challenges the claim for a certain world TW with TW =​WSp. A crucial assumption here is that the (unmarked) indicative is not a neutral mood value but identifies TW as WSp. Another crucial assumption –​or rather: principle following from Fig­ ure 1 and Figure 2 –​is that mood contributes to propositional content (also cf. Papafragou 2006: 1691–​1694). In sum, it is the modal meaning that is challenged here. B’s second reaction challenges the claim for TT, identified via its relation to TU. Put differently, what is called into question here is the restriction or anchoring of the claim to a certain time, which has been determined relative to the utterance time. Thus, the challenge targets the temporal meaning of the original utterance. B’s third reaction works the same way, except that TT is not identified via its relation to TU but via its relation to TSit; so the challenge targets the restriction necessarily brought about by the

98

Tabea Reiner

aspectual meaning. Though, it should be noted that the two reactions exploit different properties of the German [AUX +​PTCP] construction: the former picks out its temporal reading ʻTT before TUʼ, i.e. a simple past reading (Klein 1994: 126) while the latter picks out its aspectual reading ʻTT after TSitʼ, i.e. a present perfect reading (Klein 1994: 128–​129). Summarizing the results from examples (12) and (13), E patterns differently from T, A, and M with respect to challengeability, as predicted by Figure 2. However, a reviewer remarks that evidential mean­ ing might be challengeable after all if example (12) is modified in the following way: (14)

German (modification of example (12) above) A: Die Gärtnerin the female.gardener

B:

soll rep

die Juwelen gestohlen hab-​en. the jewels steal.ptcp   aux-​i nf ʻThe gardener is said to have stolen the jewels.ʼ Nein, sie will sie no she self-​r ep+​doubt them gestohlen hab-​en. steal.ptcp aux-​i nf ʻNo, she says so herself (and I doubt that she speaks the truth).ʼ

(E)

Will is the 3rd person singular form of quotative woll-​briefly mentioned above, where it was dismissed as an evidential marker for, as a quotative, being too specific. However, this does not mean that it cannot be used in a potential challenge of an evidential meaning like in (14). Its quotative function is, more precisely, one of self-​report, combined with a modal element of (severe) doubt. So, does (14) provide a counterexample to my statement that evidential meanings cannot be challenged? It does, if and only if B’s reaction can be paraphrased as ʻyour claim is not based on what you heard/​read but on a self-​incrimination of the gardenerʼ. In this case, it is truly A’s way of acquiring the relevant information that is challenged. However, this is not the way in which

Motivating the Challengeability Criterion

99

B’s reaction is usually understood. Rather, B says “As to the claim in our common ground, I do not base it on what I have heard/​read but on a doubtful self-​incrimination of the gardener”. Put differently, B says “I have a different source of information”, not “your source of information is different from the one you indicated”. I conclude that, indeed, E patterns differently from T, A, and M with respect to challengeability, as predicted by Figure 2. Thus, the demonstration announced in the title of this section has been successful. 3.5. The Inevitable Question: Are We Talking about Lexical or Grammatical Realizations of TAM and E? In connection with Anderson (1986) above I briefly mentioned Boye & Harder’s (2012) characterization of grammar vs. lexicon: grammatical elements are those whose content cannot be made at-​issue. More fully and in their own words:

Whereas lexical expressions may or may not, in actual communication, convey the main point of a linguistic message, grammatical expressions (morphemes, words, constructions) are conventionally specified as noncarriers of the main point, serving instead an ancillary communicative purpose as secondary or background elements. (Boye & Harder 2012: 6–​7)

In particular, note that the opposition is this: “may […] convey the main point” (lexical) vs. “serving […] as secondary or background elements” (grammatical). So there is no intermediate status like conveying one of two main points or conveying one of two at-​issue contents with only one of them being the real main point. Accordingly, I equate being at-​ issue with being the main point here. Against this background, recall the discussion of example (13): it was shown that the embedded claim can be challenged with respect to not only its modal but also its temporal or aspectual specification. To this extent, the modal-​temporal-​aspectual meanings involved are necessarily at-​issue (cf. note 18). Does this now mean that the pertinent expressions do not belong to grammar according to Boye & Harder (2012)? No, it does not. For it is not the modal, temporal, or aspectual meaning itself that is challenged, hence at-​issue. Rather it is the claim

100

Tabea Reiner

that is challenged for a certain world or time. In fact, one cannot challenge any anchoring or restriction as such, rather one challenges the anchored or restricted claim (recall the discussion of example (6)). That is, TAM-​meanings as such are never challengeable, hence never at-​issue (cf. note 18). By contrast, TAM-restricted claims are available for challenge and being at issue. Therefore, expressions of TAM as defined by Figure 2 of the present framework are always grammatical expressions in the sense of Boye & Harder (2012). The same holds for expressions of E as defined by Figure 2, which do not convey at-​issue content either (cf. the argu­ mentation above). However, what about evidentiality as discussed with respect to example (12)? Recall that there, on a second layer, the same type of meaning was found, though constituting a claim of its own, hence being challengeable. For the sake of clarity, I will use the following terms: • •

evidentiality as defined by Figure 2 =​ first-​order evidentiality same meaning where it additionally constitutes a claim by itself =​ second-​order evidentiality

Crucially, second-​order evidentiality is at-​issue. So its expressions, most notably finite verbs, are lexical according to Boye & Harder (2012). When Nuyts (2017) or Boye & Harder (2009) take into account also lexical expressions of evidentiality, they are, from my perspective, talking about second-​order evidentiality for a large part. Schematically: Table 1:  Scope vs. expression of evidential meaning

first-​order evidentiality second-​order evidentiality

Lexical

grammatical

-​ +​

+​ -​

The only elements that escape this classification seem to be adverbs, e.g. reportedly. They do not convey second-​order evidentiality since they do not introduce a new claim on their own but they qualify as lexical elements since their content can be at-​issue. In this sense,

Motivating the Challengeability Criterion

101

adverbs are hybrid and so it does not come as a surprise that this word class rather than any other has spawned the most heated discussions on whether or not their relevant instances (i.e. those with source-​of-​ information meaning) may count as “real” evidentials (e.g. cf. Aikhen­ vald’s 2012 review of Diewald & Smirnova 2010, in particular p. 432). Having said that, however, it is not all that easy to find or construct pertinent examples. For instance, consider the (un-​)naturalness of the following dialogue: (15)

A: B:

What is your basis for your claim that she did it? ?#She did it REPORTEDLY.

In a similar vein, (16) probably suffers from a reinterpretation of the supposed-​to-​be-​evidential adverb as a manner adverb. (16) He lied OBVIOUSLY.

The only clear example I am aware of is (17), which I overheard repeatedly in one form or another around university math classes. (17)

German (overheard) ANSCHAULICH ist das klar, aber nicht by visualization is this clear but not bewiesen. proven ʻThis (=​the truth of a given statement) is obvious when we visualize what the statement says but this does not amount to a proof.ʼ

Otherwise, examples seem to be rare. If this impression is accurate, then most adverbs with source-​of-​information meaning do fit into Table 1, after all: their meaning simply represents first-​order evidential­ ity and cannot be made at issue. That is, they are, in fact, grammatical elements in the sense of Boye & Harder (2012). This is also the conclu­ sion drawn by Boye & Harder (2009, section 4).24

24 Already the earlier publication introduces the conception of grammar vs. lexicon that is presented at length in the later one.

102

Tabea Reiner

To stay with Boye & Harder (2009) for a moment, this publica­­ tion includes as an argument for a holistic view on information source the observation that there are diachronic connections between conceptually evidential phenomena that are termed lexical and those that are termed grammatical; most notably grammaticalization processes (Boye & Harder 2009: 17–​23). However, these developments can be accounted for just as well if we say that expressions of second-​order evidentiality become expressions of first-​ order evidentiality. This acknowledges the difference in scope without neglecting the conceptual identity. And those working in an Aikhenvaldian spirit may say that proto-​evidentials become true evidentials. Generally, I am under the impression that Boye & Harder do not emphasize sufficiently what is really practical about their definition of grammar vs. lexicon: they provide two clear-​cut classes while the phenomena as such may be continuous, or rather: display flexible and perhaps even multiple class membership, especially from a diachronic perspective (cf. also Boye 2018: 268). In fact, I consider Boye & Harder’s (2009/​2012) conception of grammar vs. lexicon more hands-​on than both the classical criterion of close-​class membership, as evoked by Aikhenvald (2012: 432–​433, 2018: 3–​5), and my own distinction (Reiner 2014). The reason is that both alternatives are inconvenient to apply. The former runs into problems when it excludes adverbs like allegedly or reportedly (Aikhenvald 2012: 432) from the realm of grammar on the grounds that these are open-​class items: if these are open-​class items, also ?watchedly or ?inferredly should be accepted by English native speakers –​which is far from clear and would have to be checked by cumbersome productivity tests. The latter distinction, i.e. my own one, comes with its own kind of problems. In a nutshell, it says that we should regard as grammatical (as opposed to lexical) all signs that include in their meaning a piece of information about themselves, i.e. that are self-​relational.25 In this sense, the meanings defined by Figure 2 (including first-​but not 25 In the publication referred to, I use the term self-​referential instead of self-​ relational; however the former might suggest that the notion has to do with referents vs. concepts, which is decidedly not the case.

Motivating the Challengeability Criterion

103

second-​order evidentiality) qualify as grammatical under certain conditions: they relate to the comment of the utterance and to the extent that this is equivalent to saying that ʻthey relate to the very comment conveyed by the predicate expression that also hosts the marker of the respective meaningʼ, the pertinent signs are self-​relational in the required sense. Additionally, for tense it can be argued that it relates to TU and hence to “the time point of the act of speech, i.e., of the token uttered” (Reichenbach 1947: 288). The problem, however, is that any kind of self-​relational meaning is sufficient to classify its carrier (more precisely: the sign as a whole) as belonging to grammar. So grammar becomes large while the lexicon shrinks –​in fact, it shrinks to flexible roots, which do not even include syntactic category information (Van Lier & Rijkhoff 2013). A way out is to refrain from classifying whole elements and dissect every meaning into its grammatical vs. lexical parts.26 If this strategy is adopted, we have to say that certain parts of TAM-​or E-​meanings are grammatical and to this extent are their carriers. Doing so would be accurate but very cumbersome. This is the reason why I do not use my own conception of lexicon vs. grammar here but rather adopt the one from Boye & Harder (2009/​2012). Concluding this section, the question whether TAM-​and E-​ markers are lexical or grammatical elements constitutes the only case in which TAM and E do not behave as two groups: markers of TAM pattern with markers of first-​order E in that they are (mostly) grammatical elements while markers of second-​order E are different in that they are lexical elements. Thus, even though the present paper is about the semantic definition of tense, aspect, and mood as opposed to evidentiality, the markers of the four categories mostly turned out to be grammatical elements.

26 The lexical parts are the ones without information about the sign itself, e.g. the mental image evoked by tree but also by grows can be described without any reference to the sign as such –​whereas we cannot describe in such a way the concept evoked by the -​s.

104

Tabea Reiner

4. Conclusion In the present paper, a framework was sketched, from which [-​]‌challengeable falls out naturally as a necessary (not sufficient) criterion for evidential meanings: unlike temporal, modal, or aspectual meanings, they just do not operate on claims. However, as soon as we look beyond one simple topic-​comment structure, evidential meanings may constitute claims on their own. So there are two levels of evidentiality, one not drawing on claims, the other one consisting of claims. These two levels of evidentiality were then related to the grammatical vs. lexical status of their markers. It is my hope that the distinction will be used as a new way to talk about what constitutes a “real” evidential.

References Aikhenvald, Alexandra Y. 2012. Gabriele Diewald, Elena Smirnova (eds). 2010. Linguistic realization of evidentiality in European languages. Berlin: De Gruyter Mouton. Reviewed by Alexandra Y. Aikhenvald (The Cairns Institute, James Cook University). Studies in Language 36/​2, 431–​439. Aikhenvald, Alexandra Y. 2018. Evidentiality: the framework. In Aikhenvald, Alexandra Y. (ed) The Oxford Handbook of Evidentiality, Oxford: Oxford University Press, 1–​43. Anderson, Lloyd B. 1986. Evidentials, paths of change, and mental maps: typologically regular asymmetries. In Chafe, Wallace /​ Nichols, Johanna (eds) Evidentiality: The Linguistic Coding of Epistemology, Norwood: Ablex, 273–​312. Ayoun, Dalila /​Celle, Agnès /​Lansari, Laure 2018. Introduction: on the gradience of TAM-​E categories. In Ayoun, Dalia /​Celle, Agnès /​Lansari, Laure (eds) Tense, Aspect, Modality, and Evidentiality: Crosslinguistic Perspectives, Amsterdam: Benjamins, 1–​18. Boye, Kasper 2012. Epistemic Meaning: A Crosslinguistic and Functional-​Cognitive Study. Berlin: De Gruyter.

Motivating the Challengeability Criterion

105

Boye, Kasper 2018. Evidentiality: the notion and the term. In Aikhenvald, Alexandra Y. (ed) The Oxford Handbook of Evidentiality, Oxford: Oxford University Press, 261–​272. Boye, Kasper /​Harder, Peter 2009. Evidentiality: linguistic categories and grammaticalization. Functions of Language 16/​1, 9–​43. Boye, Kasper /​Harder, Peter 2012. A usage-​based theory of grammatical status and grammaticalization. Language 88/​1, 1–​44. Cornillie, Bert 2009. Evidentiality and epistemic modality: on the close relationship between two different categories. Functions of Language 16/​1, 44–​62. Déchaine, Rose-​Marie /​Cook, Clare /​Muehlbauer, Jeffrey /​Waldie, Ryan 2017. (De-​)Constructing evidentiality. Lingua 186/​187, 21–​ 54. Dendale, Patrick 2019. “Non-​evidentials”: the case of verbs like s’avérer, turn out, blijken, resultar. Presentation given at the workshop Evidentials versus Non-​Evidentials (16th IPrA Meeting, Hong Kong, 13 June). Dendale, Patrick /​Tasmowski, Liliane 2001. Introduction: evidentiality and related notions. Journal of Pragmatics 33/​3, 339–​348. Diewald, Gabriele /​Smirnova, Elena 2010. Evidentiality in German: Linguistic Realization and Regularities in Grammaticalization. Berlin: De Gruyter. Faller, Martina T. 2002. Semantics and Pragmatics of Evidentials in Cuzco Quechua. PhD thesis, Stanford University. . Faller, Martina T. 2006. Evidentiality below and above speech acts. Draft. Faller, Martina T. 2017. Reportative evidentials and modal subordination. Lingua 186/​187, 55–​67. Fleck, David W. 2007. Evidentiality and double tense in Matses. Language 83/​3, 589–​614. Gergel, Remus 2001. From simple predicators to clausal functors: the English modals through time and the primitives of modality. ZAS Papers in Linguistics 24, 125–​143.

106

Tabea Reiner

Grzech, Karolina 2016. Discourse Enclitics in Tena Kichwa: A Corpus-​ Based Account of Information Structure and Epistemic Meaning. PhD Thesis, University of London. Gutzmann, Daniel /​Turgay, Katharina 2019. Secondary content: an introduction. In Gutzmann, Daniel /​Turgay, Katharina (eds) Secondary Content: The Semantics and Pragmatics of Side Issues, Leiden: Brill, 1–​25. Hinterwimmer, Stefan 2019. The Bavarian discourse particle fei as a marker of non-​at-​issueness. In Gutzmann, Daniel /​Turgay, Katharina (eds) Secondary Content: The Semantics and Pragmatics of Side Issues, Leiden: Brill, 246–​273. Hintz, Daniel J. /​Hintz, Diane M. 2017. The evidential category of mutual knowledge in Quechua. Lingua 186/​187, 88–​109. Howard, Rosaleen 2012. Shifting voices, shifting worlds: evidentiality, epistemic modality and speaker perspective in Quechua oral narrative. Pragmatics and Society 3/​2, 243–​269. Iatridou, Sabine 2000. The grammatical ingredients of counterfactuality. Linguistic Inquiry 31/​2, 231–​270. Klein, Wolfgang 1994. Time in Language [Germanic Linguistics]. London: Routledge. Klein, Wolfgang 2008. The topic situation. In Ahrenholz, Bernt /​Bredel, Ursula /​Klein, Wolfgang /​Rost-​Roth, Martina Skiba, Romuald (eds) Empirische Forschung und Theoriebildung: Beiträge aus Soziolinguistik, Gesprochene-​ Sprache-​und Zweitspracherwerbsforschung; Festschrift für Norbert Dittmar zum 65. Geburtstag, Frankfurt: Lang, 287–​305. Klein, Wolfgang 2009. Finiteness, universal grammar, and the language faculty. In Guo, Jiansheng et al. (eds) Crosslinguistic Approaches to the Psychology of Language: Research in the Tradition of Dan Isaac Slobin, New York: Psychology Press, 333–​344. Korotkova, Natasha 2020. Evidential meaning and (not-​)at-​issueness. Semantics and Pragmatics 13/​4, Early Access. Leiss, Elisabeth 2012. Epistemicity, evidentiality, and Theory of Mind (ToM). In Abraham, Werner /​Leiss, Elisabeth (eds) Modality and Theory of Mind Elements across Languages, Berlin: De Gruyter, 39–​65.

Motivating the Challengeability Criterion

107

Matthewson, Lisa 2006. Presupposition and cross-​ linguistic variation, North East Linguistic Society (NELS) 26, 63–​76. Matthewson, Lisa 2012. Evidence about evidentials: where fieldwork meets theory. In Stolterfoht, Britta /​Featherston, Sam (eds) Empirical Approaches to Linguistic Theory: Studies in Meaning and Structure, Berlin: De Gruyter, 85–​114. Molsing, Karina V. /​Ibaños, Ana Maria T. (eds) 2013. Time and TAME in Language. Newcastle upon Tyne: Cambridge Scholars. Müller, Kalle 2019a. Sentence adverbials, (non-​)at-​issueness, and orientation in German: evidence from conditionals. In Gutzmann, Daniel /​Turgay, Katharina (eds) Secondary Content: The Semantics and Pragmatics of Side Issues, Leiden: Brill, 311–​340. Müller, Kalle 2019b. Reference to evidence: towards a distinction of evidential versus epistemic meanings. Presentation given at the workshop Evidentials versus Non-​Evidentials (16th IPrA meeting, Hong Kong, 13 June). Murray, Sarah E. 2010. Evidentiality and the Structure of Speech Acts. PhD thesis, Rutgers University. Murray, Sarah E. 2017. The Semantics of Evidentials. Oxford: Oxford University Press. Nuckolls, Janis B. 2018. The interactional and cultural pragmatics of evidentiality in Pastaza Quichua. In Aikhenvald, Alexandra Y. (ed) The Oxford Handbook of Evidentiality, Oxford: Oxford University Press, 202–​221. Nuyts, Jan 2001. Epistemic Modality, Language, and Conceptualization: A Cognitive-​Pragmatic Perspective. Amsterdam: Benjamins. Nuyts, Jan 2017. Evidentiality reconsidered. In Marín Arrese, Juana Isabel /​Haßler, Gerda /​Carretero, Marta (eds) Evidentiality Revisited: Cognitive Grammar, Functional and Discourse-​Pragmatic Perspectives, Amsterdam: Benjamins, 57–​83. Palmer, Frank R. 2001. Mood and Modality. 2nd edn. Cambridge: Cambridge University Press.

108

Tabea Reiner

Papafragou, Anna 2006. Epistemic modality and truth conditions. Lingua 116/​10, 1688–​1702. Partee, Barbara H. 1973. Some structural analogies between tenses and pronouns in English, The Journal of Philosophy 70/​18, 601–​609. Ramat, Paolo 2012. Review of Gabriele Diewald & Elena Smirnova, Evidentiality in German: Linguistic realization and regularities in grammaticalization (Trends in Linguistics: Studies and Monographs 228). Berlin: De Gruyter Mouton, 2010. Linguistic Typology 16/​2, 169–​177. Reichenbach, Hans 1947. Elements of Symbolic Logic. New York: Macmillan. Reiner, Tabea 2014. Lexical and grammatical meaning: revisited. In Stalmaszczyk, Piotr (ed) Semantics and Beyond: Philosophical and Linguistic Investigations, Berlin: De Gruyter, 231–​240. Reiner, Tabea 2018. Existiert ein Infinitiv Posterior im Gegenwartsdeutschen? Habilitation thesis, LMU Munich. Reiner, Tabea 2019. Variation in non-​finiteness and temporality from a canonical perspective. In Dammel, Antje /​Schallert, Oliver (eds) Morphological Variation: Theoretical and Empirical Perspectives, Amsterdam: Benjamins, 283–​310. Ritter, Elizabeth /​Wiltschko, Martina 2009. Varieties of infl: Tense, Location, and Person. In Van Craenenbroeck, Jeroen (ed) Alternatives to cartography, Berlin: De Gruyter, 153–​202. Rizzi, Luigi /​Cinque, Guglielmo 2016. Functional categories and syntactic theory. Annual Review of Linguistics 2, 139–​163. Schenner, Mathias 2008. Double face evidentials in German: reportative ‘sollen’ and ‘wollen’ in embedded contexts. In Grønn, Atle (ed) Proceedings of SuB12, Oslo: ILOS, 552–​566. Schenner, Mathias 2010. Evidentials in complex sentences: foundational issues and data from Turkish and German. University of British Columbia Working Papers in Linguistics 28, 183–​220. Shanon, Benny 1976. On the two kinds of presuppositions in natural language. Foundations of Language 14/​2, 247–​249. Tonhauser, Judith /​Beaver, David /​Roberts, Craige Roberts /​Simons, Mandy 2013. Toward a taxonomy of projective content. Language 89/​1, 66–​109.

Motivating the Challengeability Criterion

109

Vanderbiesen, Jeroen 2018. Reportative sollen in an exclusively functional view of evidentiality. In Foolen, Ad /​De Hoop, Helen /​ Mulder, Gijs (eds) Evidence for Evidentiality, Amsterdam: Benjamins, 173–​198. Van Lier, Eva /​Rijkhoff, Jan 2013. Flexible word classes in linguistic typology and grammatical theory. In Rijkhoff, Jan /​Van Lier, Eva (eds) Flexible word classes: Typological studies of underspecified parts of speech, Oxford: Oxford University Press, 1–​30. Von Fintel, Kai 2004. Would you believe it? The king of France is back! (Presuppositions and truth-​value intuitions). In Reimer, Marga /​Bezuidenhout, Anne (eds) Descriptions and Beyond, Oxford: Oxford University Press, 315–​341. Wiemer, Björn 2018. Evidentials and epistemic modality. In Aikhenvald, Alexandra Y. (ed) The Oxford Handbook of Evidentiality, Oxford: Oxford University Press, 85–​108.

Section B Evidentiality and Modality: Descriptive Issues and Corpus-​based Studies

Nicolas Tournadre

Evidential Strategies and Hierarchies in Ladakhi: the Case of Sensory Perceptions

Abstract The grammatical evidential-​epistemic systems of the Tibetic languages are one of the richest among the World languages. These systems encode the source and access to information (sensory, inferential, authoritative1 and reported speech evidentials), as well as the speaker’s degree of certainty (epistemic modality). I will deal here with the Ladakhi (a.k.a. Ladaks) grammatical system of Leh in Central Ladakh (India). This paper will concentrate on the marking of sensory perceptions as well as on the evidential strategies and hierarchies. As we will see, the sensory markers ’DUG and RAG that play an essential role in Central Ladaks cannot only be described as “visual” and “non-​visual”. In fact, ’DUG is not exclusively visual and often includes non-​visual perceptions. However, in the case of gustatory or olfactory perceptions, the use of ’DUG is quite restricted and in some situations is simply unacceptable. Various factors, such as the hierarchy of evidential functions, temporal scope and shareable sensory experience, play an important role in the choice of evidential markers. Keywords: evidential hierarchy, epistemic modality, Tibetic, Ladakhi, sensory perception, endopathic, cognitive semantics.

1

‘Authoritative’ includes factual as well as egophoric markers.

114

Nicolas Tournadre

1. Introduction All the Tibetic languages have developed evidential and epistemic systems (hence E-​E) that are among the most complex systems in the world (see Tournadre & LaPolla 2014, Gawne & Hill 2017, Tournadre & Suzuki 2021). These systems encode the source and access to information as well as the speaker’s degree of certainty. They also take into account the ethos of the speaker. Five major categories are attested: a) sensory, b) inferential, c) authoritative, d) reported speech, e) epistemic (see e.g. Tournadre 2017). These categories may combine and include various subcategories such as visual sensory, non-​visual sensory, sensory inferential, logical inferential, etc. Other secondary categories are also attested in a number of these languages (see e.g. Tournadre & Suzuki 2021). The sensory markers play a crucial role in the E-​E systems, which constitute a significant part of the grammars in the various Tibetic languages (Gawne & Hill 2017, Tournadre & Suzuki 2021). The main sen­­ sory markers that are found in a dozen of major Tibetic languages are derived from the following Classical Tibetan auxiliaries: ‘dug (Ü, Tsang, Tö, Dzongkha, Lhoke, Sherpa, Spiti, Ladaks, Purik, Kham, Southern Mustang, Kagate, Jirel), gda’ (Kham, Hor), snang (Phenpo, Kham, Balti-​ Turtuk, Nubra), grag (Tö, Spiti, Central Ladaks, Eastern Purik, Southern Kham). In a more marginal way, one also encounters the auxiliary yod (Choča-​ngača in Bhutan) and the combination yod.gi (Amdo). It should be noted right away that these markers serve to encode the categories of sensory perception and sensory inferential. Within the inferential category, one may distinguish between sensory inferences and various types of logical inferences. The difference between “(direct) sensory” and “sensory inferential” is often conveyed through aspectual or tense differences. The (direct) sensory marking occurs in the present and past progressive aspect, whereas the sensory inferential occurs in the perfect aspect and the future tense. In order to encode sensory marking, the most frequent form in the Tibetic-​speaking area is derived from the auxiliary ‘dug. It is present in Central and Western Tibet in some dialects of Kham (in Southeastern Tibet), and the Southern Himalayas and Ladakh. In its turn, the auxiliary grag performs a specific function within the paradigm

Evidential Strategies and Hierarchies in Ladakhi

115

of sensory markers: it indicates a sensory non-​visual perception. This auxiliary and the forms that are derived from it /​ɖak/​(Kham, Tö, Sham, Purik oriental), /​rak/​(Leh), /​ak/​(Nubra) always coexist with another auxiliary, namely ‘dug or snang, which conveys a visual perception. The Tibetic languages that encode visual versus non-​ visual perceptions in their E-​E systems are a minority. They are essentially located in the Western regions of the Tibetic area: in Western Tibet (Tö Ngari) and in Ladakh, except Western Purik (Zeisler 2017). As shown in e­ xamples 1 and 2, some dialects of Kham in Southeastern Tibet also possess non-​visual sensory markers (Gesang Jumian and Gesang Yangjing 2002). (1)

dbang chen

(2)

khyod -​gi rum nang stag.ga yod.grag 2nd p-​gen pocket in walnut exist+​n vis ‘In your pocket, there are walnuts.’ [inferred by touching the pocket]

slebs -​‘ong grag pr.n. arrive-​perf+​n vis ‘Wangchen has arrived.’ [based on the sound of the steps]

2. The Sensory Paradigm in Central Ladaks I will now examine the verbal system of Ladakhi, which is locally called /​ladaks skat/​. The data2 that I present here is from the dialect of Central Ladaks spoken in and around the capital, Leh. The data was collected during the seven months of fieldwork I conducted during the winters of 2017 and 2018, and the summer of 2019. 2

The examples in this article are from my data (when it is not the case, there is a specific mention of the source). I am grateful to all the Ladakhi consultants, in particular Sonam Tundup and Rigzin. The data are presented in a Wylie transliteration of written Ladakhi (see Norman 2019). The choice to provide the examples in a transliteration of written Ladakhi (rather than in a phonological transcription) allows easy comparison with other Tibetic languages.

116

Nicolas Tournadre

As in the case of other Tibetic languages, the verbal system is particularly rich since it marks not only tense and aspect but also epistemic modality and evidentiality (see e.g. Koshal 1979, 1982, Zeisler 2017). I will concentrate in this paper on the “sensory” category, and thus, only the paradigm of sensory markers will be presented here. Outside the sensory markers, there are many other forms in the E-​E system of C. Ladaks, such as egophoric, factual and non-​sensory epistemic markers. One should note that some E-​E forms are relatively rare. Generally, epistemic markers are not as frequent as evidential markers. Table 1 presents the main sensory markers (both evidential and epistemic) in combination with the various tense-​aspect categories in Central Ladaks. Table 1:  The main sensory markers in C. Ladaks. TenseAspect

Sensory

Evidential

Evidential+Epistemic

present

visual non-visual visual non-visual visual non-visual visual non-visual visual Non-visual visual non-visual visual visual non-visual visual non-visual visual

’dug ‘a.rag ‘en ’dug ‘en.rag ’dug .pin3 ‘a.rag .pin ‘en ’dug .pin ‘en.rag .pin

thig ’dug་

continuous uncompleted past continuous past Simple past resultative perfect mirative perfect pluperfect inferential future1 Inferential future 2

3

Ø ste ’dug ste. rag tog

(< ‘dug)

ste ’dug .pin . ste. rag .pin . ces ’dug ces . rag

‘thig.rag་ ‘en.yin.thig ’dug ‘en.yin.thig.rag thig ’dug .pin thig rag . pin ‘en . yin.thig ’dug .pin ‘en.yin..thig rag . pin ste yin . thig ’dug . ste yin . thig rag ste. yod.thig’dug ste.yod.thig. rag Ø ste. yod.thig’dug .pin ste. yod.thig .rag .pin ces .yin.ces ’dug ces .yin.ces rag

- og (< ‘dug)

The suffix /pin/ is often followed by the clitic /ba/: /pin-ba/.

Evidential Strategies and Hierarchies in Ladakhi

117

This table needs a few comments. First, one can see that the opposition between visual and non-​visual extends to the entire paradigm and is present in the past, present and future. It is attested not only with the purely evidential markers but also with the markers that combine both evidential and epistemic marking. The same forms are compatible with both interrogation and negation. Second, from a morphosyntactic point of view, the classical auxiliary verbs ’ ‘dug, yod, yin and grag (the modern C. Ladaks cognate is RAG) are used now as enclitic markers. However, these auxiliaries are still used as autonomous copulative and existential verbs: ’dug, rag, yin.thig.’dug, yin.thig.rag, yin.ces.’dug, yin. ces.rag, yod.thig.rag, yod.thig.’dug.4 This morphosyntactic functioning both as enclitic markers and copulative/​existential verbs is generally attested in the modern Tibetic languages. From Table 1, we may conclude that the sensory evidential category is crucial in the E-​E system of C. Ladaks and some other neighbouring languages. This is also the case in most other Tibetic languages. However, as mentioned above, the opposition between visual and non-​visual is not pervasive in the Tibetic languages and is rather limited to some Western and Southeastern languages. Another interesting specificity of the system found in C. Ladaks is the existence of past forms indicated by the morpheme /​-​pin/​derived from the Classical Tibetan pa.yin. Most of the other Tibetic languages do not make a distinction between present and past uncompleted forms.

3. Visual versus Non-​visual As shown in Table 1, the non-​visual markers all contain the form rag /​ rak/​derived from the Classical verb grag “to resonate, to sound” and the visual markers all include the forms /​duk, ruk, uk/​derived from the Classical verb ‘dug “to sit, stay”. In a first approach, we may describe

4

This list only includes the copulative and existential verbs that convey sensory evidential (and in some cases epistemic) meanings.

118

Nicolas Tournadre

the markers ‘dug /​duk/​and RAG /​rak/​(as well as their allomorphs) as conveying a “visual perception” for the former and “non-​visual” for the latter. (3)

dri.ma

(4)

‘i yul bde.mo this village beautiful ‘This village is beautiful.’

rtsog.po rag smell bad cop+​n vis ‘There is a bad smell’ or ‘S/​he has a bad smell’

‘dug cop+​v is

In (3), the auxiliary rag /​rak/​indicates that the access to information is a non-​visual olfactory perception, whereas in (4), ‘dug /​duk/​ conveys visual information. The sentences in (5) and (6) differ only in their auxiliary. The interpretation of (5) is linked with a visual experience whereas the meaning of (6) may be inferred by the non-​visual auxiliary. (5)

‘i bu.mo bde mo ‘dug this girl beautiful cop+​v is ‘This girl is beautiful.’ [visual aesthetic evaluation]

(6)

‘i bu.mo bde.mo rag this girl beautiful cop+​n vis ‘This girl is good.’ [non-​visual evaluation in this case is often associated with a sexual connotation].5

The opposition between visual and non-​visual is largely valid but, as we will see, the selection of the auxiliaries is complex and motivated by various factors.

5

Other interpretations are also possible (P.C., 3 September 2019 Leh: Rebecca Norman, Bettina Zeisler).

Evidential Strategies and Hierarchies in Ladakhi

119

4. External Observation versus Endopathic Observation In the Tibetic languages, it is important to make a distinction between: a) External observation (by the five senses) b) Inner sensation or emotion perceived by the speaker. This latter function has been called “endopathic”.6 When the Tibetic languages make a grammatical distinction between “visual” and “non-​ visual”, the endopathic is always encoded by the “non-​visual marker”, which is perfectly understandable. In languages which do not display this distinction, the endopathic function is indicated by the marker that also encodes external observation (for example ‘dug, gda’, snang). However, the two functions differ from a syntactic point of view. The endopathic is always used with the 1st person whereas external observation occurs with 2nd and 3rd persons. Let us compare the same sentences in Common Tibetan and C. Ladaks: Endopathic function (7)

nga

(8)

nga

ltogs -​sa. rag 1st p. be hungry-​u ncmp+​n vis ‘I am hungry.’ (C. Ladaks)

grod.khog ltog -​kyi (‘dug) 1st p. stomach be hungry-​u ncmp+​sens ‘I am hungry.’ (Common Tibetan)

External observation (9)

grong.gseb

6

A term coined by Tournadre (1996).

‘di snying.rje.po ‘dug village this beautiful cop +​sens ‘This village is beautiful.’ (Common Tibetan)

120

Nicolas Tournadre

(10) ‘i yul bde.mo ‘dug this village beautiful cop+​v is ‘This village is beautiful.’ (C. Ladaks)

In examples (8) and (9) the auxiliary ‘dug has two different functions. In (8) the endopathic occurs only in the 1st person and may not be used (in declarative sentences) with the 2nd and 3rd persons. In (9) the auxiliary ‘dug is related to external observation and occurs with 2nd and 3rd persons. It is not compatible with the 1st person except in some specific cases, such as the description of dreams where the speakers see themselves. In C. Ladaks, as we can see from the examples (8) and (10), the opposition between endopathic and external observation is not only marked by syntax but also by morphology with the use of a specific “non-​visual auxiliary” /​rak/​. Figure 1 resumes the correlation between the sensory types of perception (endopathic vs external) and the category of person: SENSORY PERCEPTION Internal OBS. (endopathic) with 1st pers.

External OBS. with 2nd/3rd pers.

Figure 1.  Types of sensory perception and person correlation.

5. The Marker rag /​rak/​and its Main Functions rag

/​rak/​has two main functions: a) It indicates an external observation based on a non-​visual access to information; b) It marks an endopathic access to information.

Let us examine again the examples (3) and (7), which we reproduce here for convenience’s sake:

Evidential Strategies and Hierarchies in Ladakhi

(3)

121

dri.ma rtsog.po rag

‘There is a bad smell’ or ‘S/​he has a bad smell’ (7)

nga ltogs -​sa. rag

‘I am hungry.’ [endopathic]

Inasmuch the auxiliary ‘dug marks visual information, a more complete representation of the system may be rendered in the way illustrated in Figure 2. SENSORY INNER OBS.

OUTER OBS.

RAG

NON-VISUAL VISUAL (audit. gustat. tact. olfact.) RAG ‘DUG

Figure 2.  Endopathic, visual and non-​visual perception

Even if the scheme in Figure 2 is valid for most cases, we will see that this representation faces some issues and cannot explain the subtlety of the evidential system.

6. The Predominance of Visual Information The first problem that we face when we apply the above analysis is that some frequent statements that should appear with a non-​visual marker occur with a visual marker contrary to the expectation. This is, for example, the case in the sentence below: (11)

glu

rgyal .la

btang -​‘dug

song well lv-​u ncmp+​v is ‘(She) sings well!’ [I see her singing]

122

Nicolas Tournadre

Indeed, according to the above analysis, this situation should trigger a non-​visual marker since the evaluation of the voice depends on the auditory access and not on the vision. The preference of the visual rather than the non-​visual marker allows us to draw two conclusions. The first is that ‘dug is not a purely visual marker and may indicate non-​visual access together with visual information. Thus, the above sentence could be glossed as: ‘I see her singing and I hear her and thus can say] “she sings well” ’. It should be noted that this type of synesthetic access to information is very frequent and occurs in many ordinary situations. The speaker relies on several types of sensory information simultaneously to utter his/​her statement. The second conclusion is that visual information is predominant over other types of information. If the visual dimension is present, the speakers are likely to use ‘dug. The lack of visual information prohibits its use. Thus, if one hears the song on the radio, the sentence (11) becomes unacceptable. It has to be replaced by the following sentence: (12)

glu

rgyal .la

btang -​nga. rag

song well lv-​ncp+​n vis ‘(She) sings well!’ [hearing the song on radio]

In order to illustrate the predominance of visual information over other types of information, let us take other examples: (13)

khyi-​s

(14)

khyi-​s

mug –​‘dug dog-​erg bark-​u ncmp+​v is ‘The dogs are barking!’ [visual +​auditory] mug -​ga. rag

dog-​erg bark-​u ncmp+​n vis ‘The dogs are barking!’ [auditory]. (15)

’jam.po ‘dug (be) soft cop+​v is ‘(it) is soft!’ [visual +​tactile, touching and looking at a surface]

Evidential Strategies and Hierarchies in Ladakhi (16)

123

‘jam.po rag (be) soft cop+​n vis ‘(it) is soft! [tactile, touching in the dark].

Sentence (13) will be preferred if the speaker sees the dog barking, but (14) will become compulsory in the dark when the speaker cannot rely on visual access to information. If we integrate the above remarks, it is possible to propose a refined representation of the sensory categories (Figure 3). Examples (15) and (16) both imply touching the fabric, but (15) entails that the speaker also has visual access. Thus (15) may not be uttered in the dark. SENSORY

INNER OBS.

OUTER OBS.

RAG

NON-VISUAL VISUAL+OTHER SENSES [audit. gustat. tact. olfact.] [+visual dominance] RAG ’DUG

Figure 3.  Sensory markers and integration of the visual dominance

7. The Temporal Scope of the Visual and Non-​visual Markers Another parameter that plays a role in the E-​E system of C. Ladaks is the temporal scope. It seems that rag has a more restricted scope than ‘dug. In many cases, rag implies an immediate non-​visual perception, whereas ‘dug may be used at the time of perception but also a long time after. Let’s examine two examples: (17)

phru.gu

ngu-​‘a. rag

child cry-​u ncmp+​n vis ‘The kid is crying.’ [I hear him/​her crying]

124 (18)

Nicolas Tournadre

phru.gu

ngu-​rug 7

child cry-​u ncmp+​v is ‘The kid is crying.’ [I see and hear him/​her crying]

In (17), the statement usually refers to the moment of the perception. However, in some contexts, it may also refer to a past reference point, when the speaker perceived the kid’s voice. In terms of time frame, the use of ’dug in (18) seems much more flexible. It may occur at the time of perception but also a lot later. Moreover, the use of ’dug allows a gnomic interpretation based on sensory perception. In other words, the following statement (19) may correspond to a situation where the child is no longer crying, but the speaker has seen him/​her crying repeatedly: (19)

phru.gu

mang.po ngu-​rug child a lot cry-​u ncmp+​v is ‘The kid cries a lot’. [I have heard her/​him]

This type of generalization does not seem possible with the marker rag. Since this study is preliminary, we will need more data to confirm the above tendency, but it suffices here to say that the temporal scope of the two markers differ and is wider in the case of ‘dug.

8. Personal Sensory Experience versus Shareable Sensory Experience Another important distinction between the two markers is related to the fact that ‘dug may be used for a shared or shareable experience, whereas rag points toward the personal sensory experience of the actual speaker 7

rug

is an allomorph of ’dug .

Evidential Strategies and Hierarchies in Ladakhi

125

(or reported speaker in the case of a quotation). For example, in (20), the speaker implies that other people may also have heard various rumors. (20)

(21)

dpe.ra mang.po tshor-​‘dug 1st p talk a lot hear-​u ncmp+​v is ‘I hear a lot of things (about him)’. [as you the hearer and other people can hear] nga

tshor-​r a. rag 1st p hear-​u ncmp+​n vis ‘I can hear (him/​it) /​I hear (him/​it). nga

The marker rag does not generally involve a shared experience as shown in (21). Here, the speaker exclusively mentions his own auditory perception and does not include the addressee or other people.

9. Directly Reported Information and Hearsay Aside from its inferential meaning, the sensory perfect is often used in the Tibetic languages to convey hearsay. This is a frequent strategy to report news heard on the media or the phone. In Common Tibetan (or Standard Tibetan), the auxiliaries bzhag and ‘dug are used for this purpose. In C. Ladaks, the various forms of the sensory perfect V-​/​tok/​ or V-​/​teduk/​ and V-​/​terak/​are built respectively with the auxiliaries ‘dug and rag (see the chart). However, again the two markers differ in their meaning: (22)

bstan.‘dzin

phyi.rgyal-​la

song -​tog

pr.n abroad-​dat go-​perf+​v is ‘Stenzin has gone abroad’. [I learned it from the media or by hearsay] (23)

bstan.‘dzin

phyi.rgyal-​la song -​te. rag pr.n abroad-​dat go-​perf+​n vis ‘Stenzin has gone abroad’. [He called me on the phone and gave this information]

126

Nicolas Tournadre

It is interesting to note that, for hearsay, one again finds types of distinction similar to those mentioned earlier, namely differences between “shared experience” and “personal experience”. This shows that the diverse meanings and nuances conveyed by the markers rag and ‘dug occur in the various tenses, aspects and even modalities.8

10. Sensory Markers and Non-​intentional Situations As in most other Tibetic languages that have developed egophoric markers,9 there is an opposition between intentional meanings con­ veyed by some egophoric markers and non-​intentional meanings conveyed by sensory markers. Let us compare the examples (24) and (25): (24)

mdang

(25)

mdang

rdo.rje-​na

mnyam.po

‘dzing.mo btangs -​pin yesterday pr.n-​com with fight lv-​cmp+​egoint ‘Yesterday, I fought with Dorje’ [intentionally, possibly with premeditation, etc.] ma.ne mang.po

‘thung.ste rdo.rje-​na yesterday very many drink-​co pr.n-​com mnyam.po ‘dzing.mo btangs-​te.rag with fight lv-​cmp+​n vis ‘Yesterday, I fought with Dorje’ [because I was drunk, it was not controlled nor intentional]

Since C. Ladaks has two series of sensory markers, it is possible to use rag or ‘dug depending on the situations:

8 9

About TAM categories and interaction with evidential categories, see e.g Tour­­ nadre 2004; Zeisler 2004; Guentchéva 1996, 2016; Aikhenvald 2004. The concept ‘egophoric’ was first proposed in 1992 by Tournadre to describe an evidential category of Common Tibetan (see e.g. Tournadre 1992, 1994, 2008).

127

Evidential Strategies and Hierarchies in Ladakhi

(26)

ga.Di

bzhi

tshag.pa

hour 4 until ‘(I) have been drinking until 4 am’. (27)

ga.Di bzhi tshag.pa hour 4 until ‘(We) have been drinking until 4 am’.

‘thung -​ngen.‘dug -​ste.rag drink-​cont-​cmp+​i nf+​n vis

‘thung -​ngen.‘dug-​stog drink-​cont-​cmp+​i nf+​v is

Both (26) and (27) imply that the speaker did not entirely control the situation. The use of the inferential ste.rag insists on the personal dimension of the information, whereas the inferential (s)tog (derived from ‘dug) implies that the experience has been shared by the other persons participating in the party.

11. Gustatory versus Olfactory Perceptions On the one hand, the marker rag is always compatible with non-​visual perceptions: auditory, tactile, gustatory, olfactory and endopathic. On the other hand, we have seen that ‘dug is not exclusively visual and often includes visual as well as non-​visual perceptions. According to the above analysis, various factors (predominance of the visual information, wide temporal scope, shareable sensory experience) explain why this marker often occurs instead of rag to refer to perceptions in which sight only plays a marginal role (see Section 6). Thus, the marker ‘dug is quite standard in situations involving visual/​auditory or visual/​tactile perceptions. However, in situations involving gustatory or olfactory perceptions, the use of ‘dug may become unacceptable. This seems quite logical since many situations related to gustatory or olfactory perceptions do not involve sight directly. However, it is clear that, even for these two senses, things are quite complicated. This is particularly true for gastronomic experiences, which are often associated with visual perceptions. Indeed, from a cognitive point of

128

Nicolas Tournadre

view, taste is not only related to gustatory perceptions but also to olfactory perceptions and visual cues (see e.g. Spence 2015). This cognitive explanation may account for the compatibility of ‘dug with taste experiences: (28)

rag cop+​n vis

khar.ji zhim.po food delicious ‘This food is delicious’.

As expected, the non-​visual marker rag corresponds here to the default use. However, the marker ‘dug is also acceptable: (29)

‘dug -​ha.la food tasty cop+​v is-​tag ‘This food is delicious, isn’t it?’ khar.ji

zhim.po

Here the speaker has first tasted the dish and shares his impression with the addressee(s). The confirmation question tag -​ha.la corroborates the analysis of “shared or shareable sensory experience” (see Section 7 above). With olfactory perceptions, the non-​visual marker rag is always acceptable, whereas the visual marker ‘dug is often perceived as problematic because of the absence of visual cues. Let us consider some examples. The use of the non-​ visual marker rag is readily acceptable in (30) because the speaker perceives either the smell or the noise of the fart (or both). In (31) the speaker has an olfactive access to information and is able to perceive the nice fragrance of the flower. (30)

phin.di

(31)

dri.ma

btang -​ga. rag fart lv-​u ncmp+​n vis ‘(You) are farting!’ (the speaker only perceived a bad smell or a sound).

zhim.po

smell tasty ‘(It) has a good smell’.

rag

cop+​n vis

Evidential Strategies and Hierarchies in Ladakhi

129

However, in the same situations as (30) and (31), the use of dug is often problematic. (32)

Phin.di btang ‘dug fart lv-​u ncmp+​v is ‘(You) are farting!’

Many people would accept (32) if the speaker has observed or inferred from the face or body behavior as well as from the smell and sound that the person farted. People are usually more reluctant to accept the sentence below: (33)

?dri.ma zhim.po ‘dug-​ha.la smell tasty cop+vis+tag ‘(It) has a good smell, right?’

It is interesting to note that speakers are quite divided upon the grammatical acceptability of (33). Out of the 20 consultants I interviewed, half of them, including a majority of women, strongly rejected sentence (33) arguing that it is not possible to see a smell! The others considered that the sentence was acceptable especially when the speaker intends to share his/​her perception. It would be quite instructive to have real statistic information about the acceptability of ‘dug in the context of olfactory perception.

12. Evidential Hierarchies Based on the observations and the analyses made in the previous sections (3–​6), we can already affirm that the use of the sensory markers ‘dug and rag reflect the scale specified in Figure 4.

130

Nicolas Tournadre

Visual > auditory, tactile > gustatory > olfactory > endopathic ‘dug < ….......................................................... > rag Figure 4.  The sensory markers and the accessibility scale.

When the information access is exclusively based on visual perception, the “visual” marker ‘dug is the only option, and the “non-​ visual” marker rag is unacceptable. At the other end of the scale, the visual marker is virtually incompatible with endopathic perceptions, which are always associated with rag. However, when the situation is perceived through various senses (e.g. visual+​auditory, visual+​tactile, visual+​gustatory, visual+​gustatory+​olfactory, etc.), the situation is more complicated, and both markers may occur. Whenever there is some visual cue, the “visual” marker ‘dug may be used instead of the non-​visual marker rag. As we have seen earlier (see section 6), the use of ‘dug, which is still frequent with auditory and tactile perceptions, becomes rare with gustatory perceptions and nearly absent with olfactory perceptions. Based on the above description, it is possible to propose the following evidential hierarchy, which reflects the default preference of the speaker: Visual sensory > non-​visual sensory

Various authors have proposed hierarchies to describe evidential systems (e.g. Barnes 1984, de Haan 1998, Faller, 2002). For example, Faller has proposed the following orderings: Visual > Second [speaker] > Third [speaker] visual > auditory > other sensory > inferential (result) > inferential (reason) > assumed.

According to the author (Faller 2002: 56), the ordering relation [above] is based on directness. However, we are dealing with two different kinds of directness. In the upper part, directness is measured in terms of intervening speakers. In the lower part, directness is measured in terms of amount of inference in reaching the conclusion conveyed by the ­utterance

Evidential Strategies and Hierarchies in Ladakhi

131

The two orderings reflect the distinction between source and access to information proposed by Tournadre and LaPolla (2014). One should note that in the Tibetic languages, the quotative is not in the same slot as other evidentials and combines with all the evidentials, which may be formulated as: S° (A: x) > S1 (A: x) > S2 (A: x) with S symbolizing the various speakers (S° the actual speaker and S1, S2 reported speakers), and A, the various types of access (x) to information (sensory, inferential, etc.). If we consider the major evidential categories found in most Tibetic languages and apply the criterion of directness, we need to include the categories of endopathic and egophoric (or personal information). We then obtain the following hierarchies and may rephrase Faller’s proposition in the following way: S° (A: x) > S1 (A: x) > S2 (A: x) visual/endopathic/egophoric >non-visual > inferential (result)> inferential (reason)> assumed.

Figure 5.  Source and access hierarchy.10

The endopathic and egophoric are located on the same level as the visual sensory in terms of directness. In the Tibetic languages, the endopathic encodes an inner sensation directly perceived by the speaker (see section 5). The egophoric category encodes personal information and is in most cases used with the 1st person. It implies the speaker’s consciousness or his/​her intention, which are directly accessible only to the speaker. For the Tibetic languages, which have developed an egophoric marker, the evidential access to the speaker’s activity is not achieved through “visual perception” but through “the speaker’s awareness” and marked by an egophoric evidential (see Tournadre and LaPolla, 2014) even if visual perception is present. For example,

10 The endopathic is marked by non-​visual sensory in the languages that have developed this category but in other languages it is simply marked by the sensory marker (see Section 4).

132

Nicolas Tournadre

in example (24), “yesterday, I fought with Dorje”, the speaker knows that he fought with his friend not because he saw himself fighting (!) but because he did it intentionally. He had direct access to his intention or personal knowledge, which would, of course, not be true for the 3rd person “Yesterday, he fought with Dorje”. In that case, the visual evidential would convey direct access to information. Let us note here that the use of the label “visual” for shreds of evidence related to activities performed by 1st persons is quite frequent among many authors although the access is typically not visual! When I say, “I am talking on the phone”, I do not see myself talking: I know that I am talking on the phone! Aikhenvald (2004: 232) notes: [in Ladakhi] the visual evidential […] is not used with first-​person subjects. According to Koshal (1979: 186), ‘one cannot see himself but can see only others’. This is markedly different from most other systems-​where the visual evidential explicitly implies seeing what one, or other people, are doing.11

This is in no way specific to Ladakhi and occurs in the great majority of Tibetic languages, if not all. This is the case for example in Common Tibetan, Amdo, Northern Kham, Sherpa, Common, Dzongkha and Lhoke.

13. Conclusions This paper aimed at showing the great richness and complexity of the sensory paradigm in Ladakhi (Central Ladaks dialect). This paradigm is only a constitutive part of the evidential-​epistemic grammatical system of this language. In this preliminary study, we have seen that the sensory markers ‘dug and rag cannot only be described as “visual” and “non-​visual”. Even if this analysis is mainly valid, it is insufficient to describe many cases and does not allow rendering the complexity of

11 The italics are mine.

Evidential Strategies and Hierarchies in Ladakhi

133

this system. Other factors –​such as the hierarchy of evidential functions, temporal scope and shareable sensory experience –​play a role in the grammatical system of the verb. From the above observations, one can see that the grammatical evidential-​epistemic system of Ladakhi and the systems of many other Tibetic languages can convey very subtle psychological nuances. Given these extraordinary characteristics, it would be enlightening to make extensive scale experiments using ERP techniques and other methods used in cognitive linguistics to test the Evidential-​ Epistemic systems of the Tibetic languages, particularly in the field of sensory perceptions.

Acknowledgments I would like to thank the two anonymous reviewers for their suggestions and comments on an earlier version of the manuscript. I am also grateful to Marta Carretero and Elena Domínguez for proofreading this article.

Abbreviations cmp: completed aspect co: connective com: comitative case cont : continuous aspect cop : copulative verb dat: dative ego: egophoric egointent: ego intentional exist: existential verb gen: genitive

134

Nicolas Tournadre

nvis: non-​visual evidential lv: light verb p: person perf: perfect pr.n: proper noun sens: sensory evidential uncmp: uncompleted aspect vis: visual evidential

References Aikhenvald, Alexandra Y. 2004. Evidentiality. New York: Oxford University Press. Barnes, Janet 1984. Evidential in the Tucuya verbs. International Journal of American Linguistics 50, 255–​271. Faller, Martina 2002. Remarks on evidential hierarchies. In Beaver, David I. /​Casillas Martínez, Luis D., /​Clark, Brady Z. /​ Kaufmann, Stefan (eds.) The Construction of Meaning, Stanford: CSLI Publications, 89–​111. Gawne, Lauren /​Hill, Nathan W. (eds.) 2017. Evidential Systems of Tibetan Languages. Berlin: De Gruyter Mouton. Gesang, Jumian (sKal.bzang’Gyur.med) /​Gesang, Yangjin (sKal.bzang dByangs.can) 2002. Zangyu fangyan gailun. Beijing: Minzu ­chubanshe. Guentchéva, Zlatka 1996. L’énonciation médiatisée. Louvain-​Paris: Peeters. Guentchéva, Zlatka (ed.) 2016. Aspectuality and Temporality, Description and Theoretical Issues, Amsterdam, Philadelphia: John Benjamins. de Haan, Ferdinand 1998. The Category of Evidentiality. University of New Mexico. Unpublished ms. Koshal, Sanyukta 1979. Ladakhi Grammar. Delhi: Motilal Banarsidass.

Evidential Strategies and Hierarchies in Ladakhi

135

Koshal, Sanyukta 1982. Conversational Ladakhi. Delhi: Motilal Banarsidass. Norman, Rebecca 2019. Rang skat. A Ladakhi-​ English dictionary. Leh: Melong Publications. Spence, Charles 2015. Multisensory flavor perception. Cell 161/​1, 24–​ 35. Tournadre, Nicolas 1992. La deixis en tibétain: Quelques faits remarquables. In Danon-​Boileau, Laurent /​Morel, Mary-​Annick (eds.) La Deixis, Paris : PUF, 197–​218 Tournadre, Nicolas 1994. Personne et médiatif en tibétain, La personne. Faits de langue 3, 149–​158. Tournadre, Nicolas 1996. Comparaison des systèmes médiatifs de quatre dialectes tibétains (tibétain central, ladakhi, dzongkha et amdo). In Guentcheva, Zlatka (ed) L’énonciation médiatisée, Leuven : Peeters, 195–​213. Tournadre, Nicolas 2004. Typologie des aspects verbaux et intégration à une théorie du TAM. Bulletin de la Société de Linguistique de Paris 99/​1, 7–​68. Tournadre, Nicolas 2008. Arguments against the Concept of ‘Conjunct’/​ ‘Disjunct’ in Tibetan. Chomolangma, Demawend und Kasbek. Festschrift für Roland Bielmeier zu seinem 65/​1, 281–​308. Tournadre, Nicolas 2017. A typological sketch of evidential /​epistemic categories in the Tibetic languages. In Gawne, Lauren /​Hill, Nathan W. (eds.) Evidential Systems of Tibetan Languages, De Gruyter Mouton. Tournadre, Nicolas /​LaPolla, Randy J. 2014. Towards a new approach to evidentiality. Linguistics of the Tibeto-​Burman Area 37/​2, 240–​263. Tournadre, Nicolas /​Suzuki, Hiroyuki 2021. The Tibetic Languages: An Introduction to the Family of Languages Derived from Old Tibetan. Paris: LaCITO-​CNRS. Zeisler, Bettina 2004. Relative Tense and Aspectual Values in Tibetan Languages: A Comparative Study (Trends in Linguistics, Studies and Monographs 150). Berlin, New York: Mouton de Gruyter. Zeisler, Bettina 2017. The emergence of the Ladakhi inferential and experiential markers from a marker for admirativity (non-​ commitment): the case of ḥdug and snaŋ, Journal of South Asian Languages and Linguistics 4/​2, 259–​307.

Tanja Mortelmans

Inferential must from a Contrastive Dutch-​German Perspective: Semantics and Constructional Aspects1

Abstract Although the modal verbs moeten (Dutch) and müssen (German) have a higher relative frequency than their English counterpart must, the opposite holds for the epistemic and/​or evidential (in the following: ep./​ ev.) uses of must, i.e. ep./​ev. uses of must are considerably more frequent than ep./​ev. uses of either moeten or müssen (Mortelmans 2012). Mortelmans (2012) has argued that the higher frequency of English ep./​ ev. must can be accounted for on functional grounds. More specifically, English ep./​ev. must is argued (1) to have a wider distribution as it also expresses pure conjectures which lack a straightforward evidential basis (whereas ev./​ep. müssen and moeten typically evoke the presence of some kind of evidence) and (2) ep./​ev. must is also used for intersubjective purposes, more in particular to signal consolidation and solidarity with the addressee. In the present paper, I would like to further pursue this line of contrastive research. First, the different translation strategies used to render ep./​ev. must in Dutch and German (these involve the use of adverbs, other modal verbs and zero expression) will be looked into. Second, the ep./​ev. uses of Dutch moeten and German müssen will be addressed more closely, in order to grasp the characteristics of the ep./​ ev. use of these verbs, also in contrast to English ep./​ev. must. The ultimate goal is to arrive at a more nuanced understanding of how the conceptual domain of (epistemic) medium certainty/​indirect evidentialty (see Boye 2012) is structured in German, Dutch and English. 1

I would like to thank two anonymous reviewers for their careful reading and helpful comments on an earlier version of this paper.

138

Tanja Mortelmans

The corpus on which the analysis is based consists of a self-​ c­ompiled contrastive parallel corpus of detective novels. English, ­German and Dutch originals are taken into account as well as translations from Italian and Swedish originals. Keywords: indirect evidentiality, inference, epistemic modality, contrastive linguistics, modal verbs, must

1. Introduction In this paper, epistemic-​inferential must and its Dutch and German counterparts will be studied. The starting point is the observation that must –​when used with epistemic-​inferential meaning –​does not automatically correspond to its immediate cognates in Dutch and German, i.e. the modal verbs moeten and müssen (see Mortelmans 2012), despite the fact that the latter verbs have also developed epistemic-​ inferential uses (see e.g. Baumann 2017, Diewald 1999, Mortelmans 2010). As Mortelmans (2012) has argued on the basis of a relatively small English-​Dutch-​German parallel translation corpus, the correspondence between the three verbs seems to be most outspoken when the evidence on which the inference is based is external (in the form of sensory evidence or general world knowledge), as in the following example: the protagonist sees the car and concludes that its owner Anne must be home. (1) ENG […] but he could see Anne’s car, so she must be home. She could hardly get out without her car. (CHL2, GER) DUT […] maar hij zag Annes auto wel onder de carport staan, dus ze moest thuis zijn. GER […] aber er sah Annes Auto im Unterstand, also musste sie wohl daheim sein.

2

CHL refers to the author Charlotte Link, whose novel The Watcher is part of my corpus. The compilation of the corpus will be discussed in the section on

Inferential must from a Contrastive Perspective

139

A weaker degree of correspondence between must and moeten/​müssen obtains when must expresses (more strongly speaker-​oriented) conjectures (see Squartini 2008), which are not based on external evidence. In the following example the speaker expresses a mere assumption that is not really backed up by external evidence or sound world knowledge. In this particular case, the Dutch and German translations do not contain moeten nor müssen, but feature the epistemic adverbs waar­ schijnlijk (‘probably’) and bestimmt (‘surely’) to express the speaker’s conjecture. (2) ENG I assumed,… I thought she must have gone to Tara’s for the night. That happened once before. Just one time. (PH, ENG) DUT Ik nam aan… Ik dacht dat ze waarschijnlijk bij Tara wilde gaan logeren. Dat was één keer eerder gebeurd. Eén keer maar. GER Ich dachte, dass sie bestimmt bei Tara übernachten würde. Das hatte sie schon mal gemacht. Bei einer anderen Gelegenheit.

The present paper is conceived as a follow-​up study on Mortelmans (2012). It seeks to elaborate the former study in a number of ways: empirically, since for this paper a parallel English-​Dutch-​German translation corpus was compiled which is more than five times as large as the (rather small) corpus used in Mortelmans (2012), and theoretically, as the findings of the present study will be linked to the ongoing debate on how to describe the semantics of epistemic must (see e.g. Good­ hue 2017, Lassiter 2016, von Fintel/​Gillies 2010). Moreover, explicit attention will be drawn to constructional features related to the use of epistemic-​inferential must in English and how these features influence the correspondence patterns that exist between these three verbs.

Data Collection. Table 1 presents the various novels and the abbreviations used throughout this paper. GER refers to the fact that the original novel is written in German. I will also use DUT (for Dutch), ENG (for English), SWED (for Swedish) and IT (for Italian).

140

Tanja Mortelmans

The paper is structured as follows. First, an overview of the literature on epistemic-​inferential must will be presented, which not only focuses on studies within the functional paradigm but also includes formal semantic studies. I will also briefly consider the distinctions that have been claimed to exist between ENG must and DUT moeten /​ GER müssen. I will then present the corpus that was used and the methodology, before zooming in on the semantic and formal properties of ENG epistemic-​inferential must in the corpus. This is followed by a discussion of the main translation correspondence patterns, whereby two of these patterns will be addressed in greater detail: subclause uses of must, on the one hand, and 2nd person subject uses, on the other. Attention will also be paid to some deviant translation patterns which are taken to reflect the epistemic meaning range of must. The final section presents a summary of the main findings and tries to account for them in a uniform way.

2. Three Cognates (ENG must, DUT moeten, GER müssen): Similar, but not the Same In order to get a better grasp at the differences between the three verbs, we will first have a look at English must, before presenting a short description of Dutch moeten and German müssen. 2.1. English must: Between Epistemic Modality and Inferential Evidentiality First of all, it must be noted that the modal verb must –​just like the other core modals –​is on the decline (see Leech 2011, Close/​Aarts 2010): its text frequency has decreased dramatically in the last decades. This is particularly clear for the deontic uses, but also –​albeit to a lesser

Inferential must from a Contrastive Perspective

141

extent –​for the epistemic uses (see especially Leech 2011; Kranich/​ Gast 2015). Seeing that the modal’s epistemic uses seem to be somewhat more stable, Kranich and Gast (2015) hypothesize that the English modals are specializing to express epistemic meanings, whereas semi-​ modals replace them in their non-​epistemic uses. Although I cannot pursue this hypothesis here any further, it must be noted that in my corpus material (with data from 2005 onwards), must is clearly geared toward expressing epistemic-​ inferential meaning (it is epistemic-​ inferential in more than 80% of all occurrences, see Figure 1). In her seminal analysis of the English modals, Coates (1983: 41) describes the function of epistemic must in the following way: In its most normal usage, Epistemic MUST conveys the speaker’s confidence in the truth of what he is saying, based on a logical process of deduction3 from facts known to him (which may or may not be specified).

This description combines features of epistemic modality (“the speaker’s confidence in the truth of what he is saying”) with features of inferential evidentiality (“logical process of deduction from facts known to him”). And indeed, this mixed characterization is often found in meaning descriptions of inferential markers like must (see e.g. Van der Auwera/​Plungian (1998); Squartini (2016); von Fintel/​Gillies (2010); Lassiter (2016)) –​although there seems to be a consensus to view must first and foremost as a marker of epistemic modality (expressing the speaker’s commitment in terms of high certitude, knowledge and so on) and to treat the evidential component as somehow secondary, for instance in terms of a presupposition or quantity implicature (see von Fintel/​Gillies 2010; on the general connection between evidentiality and epistemic modality, see Wiemer 2018; Boye 2012). One of the central topics in the formal semantic literature on must pertains to the fact that when using epistemic must, the speaker makes a weaker claim than when uttering an unmodalized assertion, a 3

In fact, Coates (1983) distinguishes between core (confident inference) and peripheral (logical necessity) meanings: i. confident inference: ‘I confidently infer that x’ and ii. logical necessity: ‘In light of what is known, it is necessarily the case that x’.

142

Tanja Mortelmans

phenomenon first described by Karttunen (1972). Karttunen compares (1) John must have left with (2) John has left and concludes that [a]‌man who has seen John leave or has read about it in the newspaper would not ordinarily assert (1), since he is in the position to make the stronger claim in (2). (Karttunen 1972: 12)

Several proposals are found in the (formal semantic) literature to account for this. An influential account is the one presented by von Fin­ tel and Gillies (2010), who argue that it is the (evidential) indirectness component associated with must that is responsible for the impression that with must, the speaker does not have to fully commit herself to the truth of the proposition. On their view, however, must is always strong, i.e. the speaker cannot use must to distance herself from what she is saying. Their analysis is questioned by Lassiter (2014, 2016), who argues that epistemic must does not necessarily express knowledge of the proposition nor should it necessarily be associated with a high degree of speaker certainty. Lassiter provides a number of authentic examples (see 3–​5), in which must combines with elements that express either attenuation of certainty (“I’m almost certain”, “I’m almost sure”) or explicit denial of knowledge and certainty (e.g. “I don’t know for sure”, “I’m not certain”). (3)

This is not non-​stick. There must be others, but I don’t know for sure. (Lassiter 2016: 123)

(4)

This spot might be good fishing. I’ve always thought, though I haven’t seen a soul out there trying. The land must be private, I’m almost certain. (Lassiter 2016: 125)

(5)

I have an injected TB42 turbo and don’t like the current setup. There is an extra injected located in the piping from the throttle body. Must be an old DTS diesel setup but Im [sic] not certain. Why would they have added this extra injector? (Lassiter 2016: 126)

Lassiter (2014, 2016) also shows that must can be used in contexts of so-​called implicit statistical reasoning, i.e. the speaker uses must despite the (explicitly acknowledged) uncertainty of the assumptions the conclusion is based on. The use of must then indicates that what is

Inferential must from a Contrastive Perspective

143

inferred is “the best explanation of the available evidence” (Lassiter 2016: 128, original emphasis). 2.2. On Dutch moeten and German müssen Both Dutch moeten and German müssen are high-​frequency items, whose non-​ epistemic meanings are clearly prevalent. Baumann (2017: 220), for instance, notes that of all indicative uses of GER müssen in the LIMAS-​corpus, only 4.9% are epistemic (on similar findings for Dutch moeten, see e.g. Nuyts 2001: 174; Nuyts/​Byloo/​Diepeveen 2010). With respect to the epistemic-​ inferential meanings of DUT moeten and GER müssen, it is interesting to note that their semantic descriptions generally do not differ from the ones posited for must. Nuyts (2001: 174–​175), for instance, notes that the modal verb moeten has “exactly the same inferential meaning as its English cognate must”. That this statement has to be weakened to some extent, has been shown in Mortelmans (2010) and (2012). 2.3. Differences between English must and its Cognates in Dutch and German (on the Basis of Mortelmans 2010, 2012) First corpus-​based comparisons of the three verbs were conducted in Mortelmans (2010, 2012). My main findings included •

• •

the observation that the correspondence between must, moeten and müssen was, in fact, quite low, with DUT moeten and (to a somewhat lesser extent) GER müssen having a much higher text frequency than ENG must. In the contrastive corpus that was used in Mortelmans (2010), the three verbs matched each other in only 6% of all cases (Mortelmans 2010: 137); the observation that the degree of correspondence was clearly higher when only the epistemic uses were taken into account (Mortelmans 2010: 143); the fact that correspondence between epistemic must, moeten and müssen was highest when the inference expressed was based on

144

Tanja Mortelmans

evidence external to the speaker (i.e. sensory evidence or world knowledge) (Mortelmans 2012), as in example (6), in which must expresses a strong inference based on world knowledge (finding a buried skeleton is an awful thing, not only because the person was a relative but also because it led to the conclusion that someone had buried the body). (6)

ENG “So finding her, it was awful, not just because it was her, but because somebody must have buried her”. (NF, ENG) DUT “Dus, toen we haar vonden, was dat vreselijk, niet alleen omdat zij het was, maar omdat iemand haar begraven moet hebben”. GER “Die Entdeckung der Knochen war schrecklich, nicht nur weil es Natalie war, sondern weil jemand sie dort vergraben haben muss”.

In those cases in which epistemic must refers to mere conjectures lacking clear evidential backup, Dutch and German seem to prefer other means to express this. A case in point is example (7), also discussed in Mortelmans (2012). Whereas epistemic must is used in the English original to express a mere assumption (“I always thought she must be dead”), the Dutch translation features the past tense of the modal verb zullen (zou) in combination with the modal particle wel ‘well’. In the German translation, must is left untranslated (daß sie tot ist ‘that she is dead’). (7)

ENG “Why did you assume that it was Natalie?” she asked. “I don’t know”, I replied. “I suppose it was just that Natalie disappeared, and I always thought she must be dead –​though Martha would never believe that […]”. (NF, ENG) DUT “Waarom ging u ervan uit dat die van Natalie waren?” vroeg ze. “Dat weet ik niet”, antwoordde ik, enigszins van mijn stuk gebracht. “Ik denk dat het gewoon kwam omdat Natalie verdwenen was en ik altijd dacht dat ze wel dood zou zijn”. GER “Wieso haben Sie gleich angenommen, daß es Natalies Knochen sind?“ fragte sie. “Ich weiß nicht”, antwortete ich etwas verblüfft. “Wahrscheinlich nur deshalb, weil Natalie verschwunden war und ich immer geglaubt habe, daß sie tot ist, obwohl Martha das nie wahrhaben wollte […]”

Inferential must from a Contrastive Perspective

145

Furthermore, Mortelmans (2012) claimed that English must has devel­ oped a special kind of intersubjective use (see Traugott 2003), in which the speaker expresses that she empathizes with the addressee, expresses her respect and sympathy for the addressee, seeks to avoid speech acts that impact on the addressee’s face etc. DUT moeten and GER müssen seem less apt to express such face-​saving meanings. So, in example (8), empathic must is simply not translated in Dutch, whereas German uses the adverb sicher ‘surely’ to render it. (8)

ENG “We believe you, Jane, and we trust you”. “Thank you, I’m sure this group must be very emotionally helpful”. (NF, ENG) DUT “Wij geloven je, Jane, en wij vertrouwen je”. “Dank je wel. Ik weet zeker dat deze groep jullie veel emotionele steun biedt”. GER “Wir glauben und vertrauen Ihnen, Jane”. “Danke, sicher ist Ihre Gruppe eine wichtige emotionale Unterstützung”.

As the corpus compiled by Mortelmans (2012) was quite small, it can­ not be completely ruled out that some of the findings were accidental rather than reflect structural differences between English must, on the one hand, and moeten/​müssen, on the other. Moreover, Mortelmans (2012) did hardly pay any attention to possible formal correlates of the differences found. The present study will try to remedy this.

3. Data Collection Data for the present paper were collected from eight detective novels which have all appeared in English, Dutch and German (translation) between 2005 and 2016, two of which were originally written in English (the ones by Nicci French (NF) and Paula Hawkins (PH)), two in Dutch (the ones by Bram Dehouck (BD) and Ester Verhoef (EV)), two in German (the ones by Charlotte Link (CHL) and Jan Costin Wagner (W)), one in Italian (Valerio Varesi, VV) and one in Swedish (Hjorth/​Rosenfeld (HJR)). Note that I did not consult the original versions of the Italian and Swedish novels but only looked at the translations. The main

146

Tanja Mortelmans

reason for also including translations from Swedish and Italian lies in the fact that they are typologically somewhat less related to English than (Westgermanic) German and Dutch. If similar tendencies can be found in the translations from Swedish and (particularly) Italian, we can safely assume that these similarities cannot be accounted for merely on the basis of interference between languages that have a highly similar modal system (as English, German and Dutch).4 Table 1.  Overview of the novels in the self-​compiled corpus. Original Language

English

Dutch

German

English

Nicci French (NF) =​Nicci Gerrard (°1957) /​Sean French (°1958) (UK) Sunday Morning Coming Down (2017) Paula Hawkins (PH) The girl on the train (2015) (UK)

Zondagochtend breekt aan Translator: Marja Borg (NL) /​Els Van Son (NL)

Blutroter Sonntag Translator: Birgit Moosmüller (D)

Het meisje in de trein Translator: Miebeth van Horn (NL) Bram Dehouck (BD) (°1973): Een zomer zonder slaap (2011) Esther Verhoef (EV) (°1968): Rendez-​ vous (2006) Voor de leeuwen Translator: Gerda Meijerink (NL)

Girl on the train. Du kennst sie nicht, aber sie kennt dich Translator: Christoph Göhler Sommer ohne Schlaf Translator: Stefanie Schäfer (D)

English

Dutch

A sleepless summer Translator: Jonathan Reeder (USA)

Dutch

Rendez-​vous Translator: Alexander Smith (USA) The Winter of the Lions Translator: Anthea Bell (UK) The watcher Translator: Stefan Tobler

German

German

4

De toeschouwer Translator: Corry Van Bree (NL)

Der Geliebte Translator: Ilja Braun (D) Jan Costin Wagner (W) (°1972) Im Winter der Löwen (2009) Charlotte Link (CHL) (° 1963) Der Beobachter (2011)

Of course, Swedish måste and English must are cognates as well. According to Aijmer (2018: 209), the most frequent correspondence to English must is indeed måste. Still, Aijmer’s (2018) analysis of a parallel Swedish-​English fiction corpus shows that epistemic måste is translated differently, i.e. not by must, in more than 30% of all cases.

147

Inferential must from a Contrastive Perspective Table 1.  Continued Original Language

English

Dutch

German

Italian: Valerio Varesi (VV) (2003): l fiume delle nebbie Swedish: Hjorth /​Rosenfeldt (HJR) (2011): Lärjungen

River of shadows 2010 Translator: Joseph Farrell (UK) The man who watched women (2012) Translator: Marlaine Delargy (UK)

Schaduwrivier 2012 Translator: Aniek Njiokiktjien (NL) De discipel (2012) Translator: Geri de Boer (NL)

Der Nebelfluss 2005 Translator: Karin Rother (D) Die Frauen, die er kannte. Translator: Ursel Allenstein (D)

For all eight novels, the starting point were the occurrences of must –​either in the original English versions (NF/​PH) or in the English translations. These occurrences were manually compared to the corresponding passages in the Dutch and German versions. In total, the corpus consists of 522 occurrences (each one of them in the three language versions English-​Dutch-​German), as in examples (9), (10) and (11) below. These examples make clear that the corresponding passages may contain the cognates of must (DUT moeten/​GER müssen as in (9)) or may exhibit quite different translation options (as in 10). In some cases, only two languages feature a form of must, whereas the third one deviates, as in example (11), where both ENG and DUT use must/​ moeten, whereas GER uses the adverb bestimmt ‘surely’. (9)

ENG “It must feel like a nightmare”, said Blackstock (NF, ENG) DUT “Dat moet een nachtmerrie voor u zijn”, zei Blackstock. GER “Es muss Ihnen vorkommen wie ein Albtraum”, meinte Blackstock.

(10) ENG When I noticed she wasn’t there, I thought she must be with him. (NF, ENG) DUT Toen ik merkte dat ze weg was, dacht ik dat ze wel bij hem zou zijn. GER Als mir irgendwann auffiel, dass sie nicht mehr da war, dachte ich, sie sei [=​present subjunctive of sein] mit ihm zusammen.

148 (11)

Tanja Mortelmans

ENG Josef sighed. Something must have caught his eye. He walked round the other side of the stall. (NF, ENG) DUT Josef slaakte een zucht. Alexei moest iets gezien hebben. Hij liep om de kraam heen. GER Josef seufzte. Bestimmt war Alexei wieder von irgendetwas besonders fasziniert.

Table 2.  Token frequency of must in the various corpus texts. must English Nicci French (NF) Paula Hawkins (PH) Dutch Bram Dehouck (BD) Esther Verhoef (EV) German Jan Costin Wagner (W) Charlotte Link (CHL) Swedish Hjorth/​Rosenfeld (HJR) Italian Valerio Varesi (VV)

n=​522 n=​72 n=​59 n=​18 n=​27 n=​41 n=​130 n=​117 n=​58

Table 2 presents an overview of the token frequency of must in the different novels. It is clear that the translations of the Dutch novels provide a relatively limited amount of tokens (only 45 in total), whereas the English translation of Charlotte Link (based on a German original) and Hjorth/​Rosenfeld (based on a Swedish original) contribute most occurrences to the corpus (130 and 117, respectively). It should be remarked that the token frequency of must is many times lower than the overall token frequency of either moeten or müssen in the corresponding texts (see Table 3).

149

Inferential must from a Contrastive Perspective Table 3.  Token frequency of must compared to token frequency of moeten and müssen (all word forms).

Nicci French Paula Hawkins Bram Dehouck Esther Verhoef Jan Costin Wagner Charlotte Link Hjorth/​Rosenfeld Valerio Varesi

must

&moeten

&müssen

n=​72 n=​59 n=​18 n=​27 n=​41 n=​130 n=​117 n=​58

n=​401 n=​480 n=​139 n=​ 315 (no e-​book) n=​657 n=​769 (no e-​book)

n=​385 n=​390 n=​128 n=​252 n=​144 n=​590 n= ​602 n=​ 180

In Nicci French’s novel, for instance, we find 72 instances of must, whereas DUT moeten occurs 401 times and GER müssen 385 times. This discrepancy is mainly due to the fact that both in Dutch and German, moeten and müssen are predominantly used with non-​ epistemic (deontic /​dynamic) meaning, for which English prefers other expressions like have (got) to or need to (see also Mortelmans 2010), as in the following examples: (12)

ENG She had to vent her annoyance at herself somehow. (CHL, GER) DUT Nu moest ze ergens naartoe met die boosheid op zichzelf. GER Sie musste irgendwohin mit ihrem Ärger über sich selbst.

(13)

ENG We need to work out how to handle this. (NF, ENG) DUT We moeten een plan van aanpak bedenken. GER Wir müssen uns überlegen, wie wir das handhaben wollen.

As I am mainly interested in epistemic-​inferential uses, though, I started from ENG must whose epistemic potential –​as we know from previous research –​is much higher than the one of DUT moeten or GER müssen. Of course, this means that I excluded all cases in which epistemic-​inferential moeten /​müssen in Dutch /​German is not translated by (or: does not correspond to) must. These cases are quite rare, but cannot be ruled out.

150

Tanja Mortelmans

In order to make sure that I was not leaving out too many instances on beforehand, I did crosschecks for two novels. I compared all 137 occurrences of DUT moeten in Bram Dehouck’s novel (BD) and all 144 occurrences of müssen in Jan Costin Wagner’s novel (W) with their correspondences in English. For BD it turns out that in only two cases, (past tense) epistemic-​inferential moeten is not translated by means of must, but corresponds to the epistemic adverb surely or epistemic-​inferential have to. These cases are therefore not in the corpus that was compiled. (14)

ENG Surely the drone deprived someone else from a good night’s sleep? (BD, DUT) DUT Er moest toch nog iemand zijn [lit. There must/​had to be someone else] die geen nachtrust gegund werd door het geruis.

(15)

ENG Deep down she knew it had to be true. (BD, DUT) DUT En ze wist ergens wel dat het waar moest zijn.

By contrast, there are eight instances of epistemic must in the English translation (and therefore in the corpus) that do not correspond to moeten in the Dutch original (see ­example 16). In this case, the translator has used epistemic-​inferential must without being “primed” by the original text to do so. The original passage in Dutch does not contain moeten, but uses a plain simple past tense (dat was het) which is translated by means of must. Note that the German translation does not contain müssen. (16) ENG On the way home she fantasized about what had worn Herman out. It couldn’t be Claire’s lust. She smiled. Liquor, that must be it. (BD, DUT) DUT Op weg naar huis fantaseerde ze over wat Herman zo had uitgeput. De hartstocht van Claire kon het niet zijn. Ze glimlachte. Drank, dat was het [lit. ‘Liquor, that was it’]. GER Auf dem Nachhauseweg grübelte sie darüber nach, wodurch Herman wohl so erschöpft war. Claire Leidenschaft konnte es nicht sein. Magda lächelte. Alkohol, das war’s! [lit. ‘Liquor, that was it’]

Inferential must from a Contrastive Perspective

151

In the case of Jan-​ Costin Wagner’s text (W), 8 out of 144 instances of müssen in the German original were epistemic-​inferential. They corresponded to either epistemic should and ought (5 cases), have to (1), surely (1) and zero (1), as in the following examples: (17) ENG The tickets should have reached you by now. (W, GER) GER Die Tickets müssten ihnen zugegangen sein (18) ENG Surely he could remember the people he had talked to yesterday evening? (W, GER) GER Er musste sich doch an die Leute erinnern können, mit denen er am Vorabend gesprochen hatte.

Taking the opposite perspective and starting from the English translation, we observe that must is used in 10 cases (out of 28 epistemic-​ evidential occurrences in this text) in which the German original did not feature a form of müssen. Again, the translator is not primed by the German original to use must, but chooses must despite the fact that the German original contains something else. (19) ENG He had no idea what this woman must be feeling […] GER Er hatte keine Ahnung, was diese Frau empfand […] (W, GER) (20) ENG “Olli Latvala”, he said. “You must be the gentleman from the police?” (W, GER) GER “Olli Latvala”, sagte er. “Sie sind der Herr von der Polizei?”

It seems therefore safe to conclude that epistemic must clearly outnumbers epistemic moeten/​müssen (even if we do not take every epistemic-​inferential instance of moeten/​müssen into account) so that it is methodologically sound to start from English must. In the following section, we will analyze the behavior of epistemic-​inferential must and the various translation options. We will first present the more general findings with respect to must and the main translation correspondence patterns, before we address two phenomena in greater detail: subclause uses and 2nd person uses of must. Finally, we will also address the epistemic strength of must on the basis of the translation patterns.

152

Tanja Mortelmans

4. General Findings 4.1. The Characteristics of must As expected, the epistemic use of must is dominant in all texts. In total, epistemic-​inferential uses account for 422 instances of must (about 81%), whereas only 100 (or about 19%) of all occurrences are found with non-​ epistemic meaning.

Figure 1.  Epistemic vs. deontic uses of must.

Quite strikingly, the majority of epistemic-​ inferential uses of must (55%) patterns with a perfect infinitive (234/​422 epistemic instances). It follows that English epistemic-​inferential must most typically expresses an inference with respect to a past event.5 The basis 5

This observation is by no means superficial. The German epistemic modal dürfte, for instance, hardly combines with a perfect infinitive (cf. Mortelmans 2019), which can be accounted for by the fact that its inferential semantics is of a completely different nature. By means of dürfte the speaker typically refers to the common ground and uses this common ground to achieve a conclusion that is shared by the speech participants.

Inferential must from a Contrastive Perspective

153

for the inference (i.e. the evidence) lies in the present, though, often in (the observation of) a particular result. On observing this result (for instance, a door being open, a flat battery, the light being dim, a particular smell) the speaker infers that something has occurred or was the case (i.e. the proposition p). (21)

It was empty. The door was open, so she must just have gone out. He hid the knife in the palm of his hand […] (HJR, SWED)

(22) Soneri took out his telephone and saw it had no battery left. He must have switched it on by mistake. (VV, IT) (23)

The light was dim. David must have covered the window. (NF, ENG)

(24) She must have bled a great deal, Sebastian thought, if the smell was still around. (HJR, SWED)

It should be noted that in most of the cases, the speaker’s inference is based on (somewhat less concrete, less directly accessible) world knowledge. In example (25), for instance, the fact that a child was abused leads the speaker to the inference that the abuse must have affected the child’s schoolwork. Or in (26), a connection is made between having an affair and the inference that a person must have been unhappy in her relation. In (27), the speaker refers to the addressee’s father who died very young and infers that this must have been tough for the addressee. (25)

“Didn’t anybody find out? Didn’t anybody suspect anything” Billy was leaning forward, interested. “I mean, it must have affected his schoolwork, among other things”. (HJR, SWED)

(26)

But if she was having an affair, she must have been unhappy, mustn’t she? (PH, ENG)

(27)

It must have been very tough when he died so young, said Gillian. (CHL, GER)

The speaker can make the inferencing process –​the link between the evidence and his or her reasoning –​more explicit by means of causal subclauses introduced by because, or by consecutive adverbs

154

Tanja Mortelmans

like therefore or (and) so or by embedding must in the main clause of a deductive if-​then statement. (28)

My handbag is not on the floor, it’s not hanging over the back of the chair where I usually leave it. I must have had it, though, because I’m in the house, which means I have my keys. (PH, ENG)

(29)

“He wasn’t seeing her and so he must have assumed she had gone off”, suspected Christy. (CHL, GER)

(30)

If Carla Roberts heard the lift unnaturally often on her floor, then someone must have sent it up there. Or gone up in it. (CHL, GER)

These explicit signals, however, are not highly frequent. In fact, they appear in only 28 of 422 (i.e. in less than 7%) epistemic-​inferential instances. Apart from the perfect infinitive, epistemic must is typically followed by stative verbs, among which be is the most frequent one. Table 4.  Infinitives with epistemic-​inferential must. epistemic must must +​perfect infinitive must +​ be must +​ have must +​ look must +​ feel must +​ know must +​ sound must +​ see must +​ take must +​other inf (make, mean, own, remember, run…)

n=​422 234 120 12 8 7 7 3 2 2 27

As far as the person of the subject is concerned, epistemic-​ inferential must occurs with first, second and third person subjects in the corpus, whereby third person subjects clearly dominate.

Inferential must from a Contrastive Perspective

155

Table 5.  Subjects with epistemic-​inferential must in the corpus. epistemic must 1st person subject 2nd person subject 3rd person subject

n=​422 22 38 362

The presence of first person subject uses (in 5% of all cases in the corpus) might be unexpected: with a first person subject, the speaker draws a conclusion with respect to a proposition she herself participates in. More than half of all first person subjects are found in Paula Hawkins’s novel (see ­examples 31 to 33), in which the protagonist has a drinking problem so that she cannot remember what she has done or what has happened after she passed out. It should be noted that inferences in this context cannot always be taken to imply knowledge or (a high degree of) certainty on the part of the speaker who merely seems to be guessing what has happened. I will come back to this issue in a later section. (31) I sliced through the top of my finger while chopping the onions. I must have gone to the bathroom to clean it up and gone to lie down for a while and just forgotten all about the kitchen, because I woke up around ten and I could hear Cathy and Damien talking and he was saying how disgusting it was that I would leave the place like that. (PH, ENG) (32) I must have fallen asleep, the gin and the hot sun lulling me. I woke with a start. (PH, ENG) (33) It had to be something to do with Saturday night. I must have done something. I must have committed some terrible act and blacked it out. (PH, ENG)

In combination with a second person subject (n=​38), the speaker infers something about the addressee, typically about his or her mental or psychological state. Note that these second person uses behave somewhat deviantly compared to the “normal” use of epistemic-​inferential must. For one thing, epistemic-​inferential [you must y] does not typically combine with a perfect infinitive (in only 11 out of 38 cases). And for another, verbs other than be ( feel, miss, remember) that are generally rather infrequent in the corpus, occur remarkably often in

156

Tanja Mortelmans

combination with a second person subject (in 11 of 26 cases), as is illustrated by the examples below. (34)

“And yet you must feel vindicated. In a way”. (NF, ENG)

(35)

“You must miss them [=​the addressee’s children]”, she said. (NF, ENG)

(36)

“You must be feeling completely dreadful. It’s unreal”. (NF, ENG)

(37)

“I know what you must feel”, said Petra. (NF, ENG)

(38)

“I’d better go. You must have a lot to do”. (CHL, GER)

(39)

“You slept”, he says again. “You must be feeling better”. (PH, ENG)

(40)

“Here, something to read. You must be bored to death, right?” (CHL, GER)

(41)

“We’re done. You must be exhausted”. (NF, ENG)

These deviant formal properties could be taken to suggest that [you must y] presents something of a special case, also from a semantic-​ pragmatic point of view. Indeed, these uses are intersubjective in that the speaker expresses “attention to the ‘self’ of the addressee/​reader in both an epistemic sense (paying attention to their presumed attitudes to the content of what is said), and in a more social sense (paying attention to their ‘face’ or ‘image needs’ associated with social stance and identity)” (Traugott 2003: 128). Such empathic uses are typical of English must (Mortelmans 2012), but occur less often with Dutch moeten or German müssen, as will be shown later on. A final observation pertains to the fact that epistemic-​inferential must clearly prefers the main clause. In 340 of 422 epistemic cases (80.6%) epistemic-​inferential must occurs in the main clause, whereas subclause uses of epistemic must account for about 19% of all instances of must, as in the following examples (42–​44): (42)

Right now I could understand better than ever how unbelievably difficult it must have been for them. (EV, DUT)

(43)

He did not know what he would find beyond that hill, but from the sheep’s excited trotting he concluded that it must be paradise. (BD, DUT)

Inferential must from a Contrastive Perspective

(44)

157

She did not know how much time had passed, but she guessed that it must have been a good three quarters of an hour before she had worked it free of the glue. (CHL, GER)

Let’s take stock: a central or prototypical use of epistemic-​ inferential must can be identified on the basis of the corpus data –​ here, epistemic-​inferential must occurs most typically in a main clause combined with a perfect infinitive or be and with a 3rd person subject. Semantically, it expresses an inference with respect to a past event based on world knowledge (or also, but less often, on sensory evidence), as in (45): (45) The light was dim. David must have covered the window. (NF, ENG)

In the following section, we will address the various translation options and try to establish whether there is a link between this core use, possible less central (i.e. less frequent) uses and the translation options that are found.

Figure 2.  Frequencies of epistemic must, moeten and müssen.

158

Tanja Mortelmans

4.2. Main Translation Correlations If we look at the correspondences of epistemic must, it is striking that in most texts (with the exception of EV), epistemic must occurs almost twice as often as its “pendants” in German and Dutch: epistemic must (n=​422) =​ DUT moeten (n=​192) /​GER müssen (n=​237). Of course, I am well-​aware that some cases have been missed (see above), in which epistemic moeten or müssen is used in Dutch or German but does not correspond to epistemic must (and is therefore not in the corpus) but these occurrences hardly make up for the remarkably high frequency of epistemic must compared to moeten and müssen. The frequency difference between DUT moeten and GER müssen –​whereby GER müssen occurs somewhat more frequently than DUT moeten –​is mainly due to one text (the German novel by Charlotte Link), where the Dutch translator seems to have avoided moeten as a translation equivalent for müssen. There were 76 instances of müssen in CHL which corresponded to moeten in only 38 cases. Quite often (26 times), the modal was not translated at all, as in the following examples. Note that I do not have a good explanation for this difference. (46)

GER Sie sah, dass Tara zusammenzuckte, und begriff, wie heftig ihr Nein geklungen haben musste. [lit. ‘must have sounded’] (CHL, GER) DUT Ze zag dat Tara in elkaar kromp en begreep dat haar ‘nee’ heel heftig had geklonken [lit. ‘had sounded’].

(47)

GER Dann geht das Licht aus. Das Auto muss irgendwo stehen [lit. ‘the car must be there somewhere’]. Und ich weiß nicht, was der Fahrer will. (CHL, GER) DUT Dan gaan de koplampen uit. De auto staat ergens [lit. ‘the car is there somewhere’] en ik weet niet wat de chauffeur wil.

In the other German original (W), the correspondence between GER müssen and DUT moeten is much higher: out of 18 uses of müssen, 16 correspond to moeten. If we only take the original English texts into account (Nicci French /​Paula Hawkins) and look at how epistemic must is translated, it becomes clear that Dutch moeten and German müssen are quite similar, at least as far as their frequency is concerned: there are 113 occurrences of epistemic must, 45 of which correspond to DUT moeten and 42 to GER müssen.

159

Inferential must from a Contrastive Perspective

If we then try to establish correlation patterns, it becomes clear that the correlation between must and its German and Dutch cognates is highest in the central cases, i.e. when epistemic-​inferential must combines with a perfect infinitive or be, appears in a main clause and the subject is 3rd person. Cases in point are the following ones (48–​50); there are in total 148 occurrences in which the three verbs fully match. (48)

ENG He stepped into the kitchen, which was no warmer than outside. The door must have been open for ages. (CHL, GER) DUT Luke liep de keuken in, waar het niet warmer was dan buiten. De deur moest al een eeuwigheid openstaan. GER […] Die Tür musste schon eine halbe Ewigkeit offen stehen.

(49)

ENG And he’d followed me, this morning. He must have been waiting outside here for at least an hour. (EV, DUT) DUT En hij was me gevolgd, vanochtend. Hij moest al minstens een uur hier buiten hebben staan wachten. GER Und jetzt verfolgte er mich offenbar auch noch. Er musste mindestens eine halbe Stunde hier draußen gewartet haben.

(50)

ENG “But if she was having an affair, she must have been unhappy, mustn’t she?” (PH, ENG) DUT “Maar als ze een minnaar had, dan moet ze wel ongelukkig zijn geweest, hé?” GER “Aber wenn sie eine Affäre hatte, muss sie doch unglücklich gewesen sein, oder nicht?”

Table 6.  Correlation between must and moeten/​müssen with infinitive as independent variable.

must =​DUT moeten must =​GER müssen

perfect infinitive (n=​234)

non-​perfect infinitive =​ be (n=​105)

non-​perfect infinitive other than be (n=​83)

119 (=​50.8%) 139 (=​59%)

45 (=​42.8%) 57 (=​54.2%)

28 (=​33.7%) 38 (=​45.8%)

160

Tanja Mortelmans

Table 7.  Correlation between must and moeten/​müssen with person of the subject as independent variable.

must =​DUT moeten must =​GER müssen

3rd person (n=​362)

2nd person (n=​38)

1st person (n=​22)

172 (47.5%) 208 (57.5%)

12 (31.6%) 15 (39.5%)

9 (40.9%) 8 (36.4%)

Table 8.  Correlation between must and moeten/​müssen with main clause/​subclause use as independent variable.

must =​DUT moeten must =​GER müssen

main clause (n=​340)

subclause (n=​80)

163 (47.9%) 197 (57.9%)

28 (35%) 39 (48.9%)

Tables 6, 7 and 8 show that the correlation goes down in those cases which do not fulfill one of these conditions. When combining some of these parameters, the correlation becomes even smaller. So, for instance, if we only take the 2nd person subjects combined with non-​perfect infinitives other than be into account (n=​15), we find that only 4 of them correspond to moeten in Dutch and 5 to müssen in German. Combining subclause uses with non-​perfect infinitives also yields a diminished degree of correlation (albeit mainly in Dutch): out of 39 relevant cases, only 10 (=​25%) correspond to moeten and 18 (=​46%) to müssen. Of course, these figures do not explain why this should be the case. Therefore, we want to have a closer look at the data to find an explanation for the fact that e.g. epistemic-​inferential must in subclauses tends not to be expressed by moeten (and to a somewhat lesser extent) müssen. We will focus on two phenomena: subclause uses, on the one hand, and the 2nd person uses, on the other.

Inferential must from a Contrastive Perspective

161

4.3. Subclause Uses Table 8 above has indicated that the subclause use of must corresponds to moeten and müssen to a somewhat diminished degree when compared to the general main clause uses. Note that the category of subclause uses includes relative clauses, concessive clauses and the like. In this section, however, I want to focus on cases like the following in which must patterns in a complement clause governed by verbs like think, assume, conclude, realize, cross her mind, reveal, see or seem. The following examples illustrate the patterns that occur: in (51) and (52), English must does not correspond to its cognates in German and Dutch. (51)

ENG “Put it in the hallway”, she said from what he suspected must be the bathroom. (HJR, SWED) DUT “Zet hem maar in de hal,” zei ze vanuit wat hij vermoedde dat de badkamer was. [lit. ‘which he suspected that it was the bathroom’] GER Stell ihn in den Flur, rief sie aus einem Raum, der vermutlich das Bad war. [lit. ‘which supposedly was the bath room’]

(52)

ENG “I assume the whole thing must be upsetting for you. After all, you knew him”. (NF, ENG) DUT “Ik neem aan dat het voor u een klap is [‘lit. ‘that it is a blow for you’]. Per slot van rekening kende u hem”. GER “Ich nehme an, Ihnen setzt das Ganze auch sehr zu. Schließlich kannten Sie ihn”.

In (53) and (54), on the other hand, we see that must equals DUT moeten /​ GER müssen. (53)

ENG John checked quickly to see which way was south and realized that in the summer, the trees must swallow up all the sunlight. (CHL, GER) DUT John [...] stelde vast dat ze vooral in de zomer het licht en de zon volkomen moesten tegenhouden GER […] und stellte fest, dass die Bäume vor allem im Sommer Licht und Sonne fast vollkommen schlucken mussten.

162 (54)

Tanja Mortelmans

ENG I wake and I know the dream must come from an old memory, some ancient transgression –​it doesn’t matter which one now. (PH, ENG) DUT […] en ik wist dat de droom afkomstig moest zijn uit een oude herinnering, een misstap uit een ver verleden –​het maakt nu niet meer uit welke. GER Als ich aufwache, ist mir klar, dass dieser Traum auf einer alten Erinnerung beruhen muss, einem längst vergangenen Fehltritt –​welchem, ist mittlerweile fast schon gleichgültig.

Table 9.  Strong epistemic-​inferential must in subclauses. certain, inferential

=​ DUT moeten

=​ GER müssen

realize that must p (n=​9) know that must p (n=​4) understand that must p (n=​3) mean that must p (n=​2) conclude that must p (n=​1) Total number: n=​19

3 3 2 2 1 =​ 11

7 2 3 1 1 =​ 14

Table 10.  Weak epistemic-​inferential must in subclauses. less certain, no/​weakly inferential

=​ DUT moeten

=​ GER müssen

think that must p (n=​14) assume that must p (n=​2) guess that must p (n=​4) suspect that must p (n=​1) suppose that must p (n=​1) (no) idea that must p (n=​4) Total number: n=​26

2 0 1 0 0 1 =​ 4

5 0 0 0 0 1 =​ 6

A closer look at the subclause uses reveals that a distinction can be made between those cases in which the semantics of the higher verb is in line with the traditional “strong” inferential semantics of must (e.g. know, realize, see Table 9) and such cases in which the higher verb codes a weaker epistemic judgment which does not

Inferential must from a Contrastive Perspective

163

automatically imply an inference by the speaker (e.g. think, assume, suppose, see Table 10).6 Strong inferential uses of must tend to have the cognate in the parallel language version; in cases in which the higher verb codes a weaker epistemic judgment, must is not so often matched by moeten/​ müssen, as can be deduced from Tables 9 and 10. What Tables 9 and 10 also show is the possible variation between German and Dutch. In the following example, for instance, the Dutch translator has opted for a sentence without moeten –​in line with our general findings –​whereas the German version has a form of müssen. What is remarkable here is that the main verb is not glauben ‘think’. Instead, we find a construction with “sein erster Eindruck war” (lit. “his first impression was”) which has a stronger inferential touch than the verb glauben ‘think’, which would be the immediate counterpart of think. (55) ENG “The autopsy is being done now, but Ian thought he must have been dead four of five days” (NF, ENG) DUT “Ze zijn nu bezig met de autopsie, maar Ian denkt dat hij al vier à vijf dagen dood is”. GER Die Autopsie ist noch nicht abgeschlossen, aber Ians erster Eindruck war, dass Stringer schon vor vier oder fünf Tagen gestorben sein muss.

4.4. 2nd Person Subjects The presence of 2nd person subjects also diminishes the likelihood that epistemic-​inferential must corresponds to DUT moeten or GER müssen (see Table 7 above). We have already noticed that 2nd person uses of must deviate from the “normal” pattern in that combinations with a perfect infinitive are quite rare. Instead, other verbs like miss, feel, remember (that are generally infrequent in the corpus) occur remarkably often. Typically, these verbs refer to emotional states or mental activities experienced by the addressee, which are not immediately accessible to the speaker. 6

In the formal semantic literature, the opposition is tied to the distinction between factive and non-​factive cognitive attitude predicates (see Mihoc et al. 2019).

164

Tanja Mortelmans

If we look at the German and Dutch correspondences of the sentences listed in the following examples, it becomes immediately clear that moeten and müssen are not the preferred correlates for must, although they do occur. Instead, we typically find a wide range of alternatives: DUT zullen ‘will’, epistemic adverbs like DUT vast ‘surely’, DUT sicher ‘surely’ and GER bestimmt ‘surely’, evidential adverbs like GER offenbar ‘apparently’, modal particles like DUT wel ‘probably’ or a zero or completely different translation (e.g. 56 GER, 59 DUT/​GER, 62 GER). (56)

ENG “And yet you must feel vindicated. In a way”. (NF, ENG) DUT “Maar het zal [‘will’] toch ook als een zekere genoegdoening voelen”. GER “Und trotzdem eine Bestätigung”. [lit. ‘and nevertheless a confirmation’]

(57)

ENG “You must miss them [=​the addressee’s children]”, she said. (NF, ENG) DUT “Je zult ze wel [‘will … probably’] vreselijk missen”, zei ze. GER “Bestimmt [‘surely’] fehlen dir die beiden sehr”, meinte sie.

(58)

ENG “You must be feeling completely dreadful. It’s unreal”. (NF, ENG) DUT “Dit moet verschrikkelijk voor u zijn. Onwerkelijk”. GER “Bestimmt [‘surely’] fühlen Sie sich absolut schrecklich. Das Ganze ist so irreal”.

(59)

ENG “I know what you must feel”, said Petra. (NF, ENG) DUT “Ik snap dat je dat denkt” [lit. ‘that you think that’], zei Petra, […]. GER “Ich weiß, wie Sie sich fühlen” [lit. ‘how you are feeling’], meinte Petra.

(60)

ENG “I’d better go. You must have a lot to do”. (CHL, GER) DUT “Dan kan ik nu beter gaan. Je hebt vast [‘surely’] veel te doen”. GER “Ich gehe dann jetzt besser. Du hast sicher [‘surely’] eine Menge zu tun”.

(61)

ENG “You slept”, he says again. “You must be feeling better”. (PH, ENG) DUT “Je sliep”, zegt hij nogmaals. “Dan voel je je vast [‘surely’] beter”. GER “Du hast durchgeschlafen”, wiederholt er. “Offenbar [‘apparently’] geht es dir besser.”

(62)

ENG “Here, something to read. You must be bored to death, right?” (CHL, GER) DUT “Hier. Iets te lezen. Je verveelt je natuurlijk dood”. [lit. ‘you are of course bored to death’] GER “Hier. Etwas zum Lesen. Sie langweilen sich zu Tode, stimmt’s?” [lit. ‘you are bored to death, is that correct?’]

Inferential must from a Contrastive Perspective

(63)

165

ENG “We’re done. You must be exhausted”. (NF, ENG) DUT “We zijn klaar. U bent vast [‘surely’] doodop”. GER “Wir sind fertig. Bestimmt [‘surely’] sind Sie erschöpft”.

Examples (56–​63) can be taken to feature intersubjective uses of must, in which the speaker uses must to express her empathy with the addressee, who is assumed to suffer or have suffered to a more or lesser degree. Quite strikingly, German7 and Dutch seem to avoid the use of müssen and moeten in this context, which suggests that the “empathic” potential of these latter verbs, i.e. the ability to express meanings having to do with (saving/​strengthening) the face of the addressee, is less developed8 than the one of their English counterpart. On this interpretation,

7 Interestingly, Letnes (2010) describes so-​called ‘affective’ uses of German epis­­ temic müssen and explicitly notes that some of them express empathy, as in: (i) Armer Kerl, dir müssens’s aber gehörig zugesetzt haben (Zweig: 317, quoted from Letnes 2010: 155) [‘Poor guy, they must have given you a hard time’, my translation] However, in the majority of examples presented in Letnes (2010), these affec­ tive uses of epistemic müssen seem to be more concerned with the speaker than oriented toward the face or social identity needs of the addressee. For one thing, affective müssen is often used to assess a proposition as particularly important, serious, critical or impressive from the point of view of the speaker (see also Ulvestad 1984). Moreover, affective müssen is shown to be often used in hyper­ bolic fashion (the speaker clearly exaggerates: Du musst doch halb tot sein nach dem Getümmel und Getue am Morgen ‘You must be half dead after the hustle and bustle of the morning’ (Böll: 293, quoted from Letnes 2010: 157)) and is often found in exclamative contexts (Sie müssen doch schon ‘ne Masse erlebt haben! ‘You must have gone through a lot!’ (Fallada: 444, quoted from Letnes 2010: 160); Sie müssen verrückt sein! ‘You must be crazy!’ (Andersch: 229, quoted from Letnes 2010: 154)). 8 Note that I do not claim that müssen and moeten do not occur at all in empathic uses. Such cases, however, are quite rare; I only found 2 of them in the corpus, as in (ii) –​which –​perhaps not incidentally –​can be characterized as an exaggeration (see footnote 7). (ii) ENG “Come in. The stairwell isn’t exactly warm. You must be half frozen to death”. (CHL, GER) DUT “Het trappenhuis is ijskoud. Je moet tot op het bot verkleumd zijn”. GER “Das Treppenhaus ist ganz schön kalt; Sie müssen fast erfroren sein”.

166

Tanja Mortelmans

German and Dutch moeten/​müssen are claimed to have retained more of their original obligation semantics, which is, in fact, face threatening, in that the speaker tries to make the addressee do something, i.e. imposes her will on the addressee. The relatively rare cases in which 2nd person epistemic/​inferential must corresponds to moeten/​müssen (8 in total) are typically not face saving, but rather face threatening to the addressee, as the following examples clearly illustrate. In (64), the speaker explicitly marks his inference9 as unpleasant for the addressee (“forgive me”), whereas the content of (65) is a vehement critique of the addressee (“you must be the worst mother in the world”). (64)

ENG “We don’t know what the background is. It’s all the more important for you to remember every detail of today’s incident. You must… forgive me, but you must have noticed something”. (W, GER) DUT “U moet toch… neemt u me niet kwalijk, maar u moet toch iets hebben waargenomen”. GER “Sie müssen doch… entschuldigen Sie, aber Sie müssen doch irgendetwas wahrgenommen haben”.

(65)

ENG […] but I could see it on their faces: thinly disguised disapproval. How could you entrust your child to that monster? You must be the worst mother in the world. (PH, ENG) DUT […] maar ik zag het op hun gezicht: nauwelijks verhulde afkeuring. Hoe had ik mijn kind aan dat monster kunnen toevertrouwen? Ik moest wel de slechtste moeder ter wereld zijn. GER Du musst die schlechteste Mutter der Welt sein.

The examples in (64) and (65) clearly do not contain empathic instances of must, which might explain why they correspond to DUT moeten and GER müssen.

9

Note that the use of must in (64) also has a deontic feel, as the speaker seems to express a kind of personal obligation of the addressee to have noticed something.

Inferential must from a Contrastive Perspective

167

4.5. How Strong is Epistemic-​inferential must? At least on the basis of the translations, it becomes clear that readers (and therefore also translators) interpret the epistemic range of ENG must as quite broad. If we look at the adverb(ial)s that correspond to ENG must, we first of all find the “usual suspects”, i.e. strong epistemic adverbs like DUT vast ‘surely’ (n=​30), DUT zeker ‘surely’ (n=​5), DUT beslist ‘surely’ (n=​2), GER bestimmt ‘surely’ (n=​33), GER sicher ‘surely’ (n=​11), all of which express high certainty. Furthermore, the fact that must often corresponds to evidential adverbs makes clear that its evidential semantics is quite accessible: we find DUT blijkbaar ‘apparently’ (n=​16) as the third10 most frequent cor­ responding adverb in Dutch and GER offenbar ‘apparently’ (n=​19) as the second most frequent adverb in German. Other clearly evidential adverb(ial)s include DUT kennelijk ‘apparently’ (n=​3) and GER allem Anschein nach ‘apparently’ (n=​1), GER anscheinend ‘apparently’ (n=​2) and GER offensichtlich ‘apparently’ (n=​2). What seems to be most striking, however, is the fact that ENG must also corresponds to weak(er) epistemic markers like DUT waarschijnlijk ‘probably’ (n=​ 22), DUT vermoedelijk ‘supposedly’ (n=​ 1), 11 misschien ‘perhaps’ (n=​5), DUT zullen ‘will’ (n=​15), the weak modal verb DUT kunnen ‘can, may’ (n=​2), GER wahrscheinlich ‘probably’ (n=​9), GER vielleicht ‘perhaps’ (n=​1), GER können ‘can, may’ (n=​3), GER mögen ‘may’ (n=​2), GER vermutlich ‘supposedly’ (n=​5) and GER dürfte ‘probably’ (n=​3), as in the following examples. (66)

ENG He had dark rings under his eyes and his lips were ashen. He must have been partying well into the night. (CHL, GER) DUT Hij had donkere kringen onder zijn ogen en bleke lippen. Waarschijnlijk [‘probably’] had hij tot ‘s ochtends vroeg gefeest […].

10 The epistemic adverb vast ‘surely’ occurs most often, waarschijnlijk ‘probably’ is in second position. 11 The modal verb zullen is often (9/​15) accompanied by the particle wel. With regard to the combination ‘zal wel’, Hogeweg (2009) notes that “the speaker seems to indicate that based on the information available to him, he can draw a certain conclusion, but that he is aware of the fact that the available information is not sufficient to be certain” (Hogeweg 2009: 534).

168

Tanja Mortelmans GER Braune Ringe unter den Augen, graue Lippen. Wahrscheinlich [‘probably’] hatte er bis in die frühen Morgenstunden gefeiert […].

(67)

(68)

ENG “He’s got it into his head to cycle thirty kilometers a day. I wonder how he’ll keep it up. […] There must be a girl in the picture, heh?” (BD, DUT) DUT “Hij heeft het in zijn hoofd gehaald om elke dag dertig kilometer te fietsen. Ik vraag me af hoe hij dat gaat volhouden. … Er zal [‘will’] een meisje in het spel zijn, zeker?” GER “Er hat sich in den Kopf gesetzt, jeden Tag dreißig Kilometer Rad zu fahren. Mal sehen, ob er das wirklich durchhält. […] Da ist bestimmt [‘surely’] ein Mädchen im Spiel”. ENG The show went on. Hämäläinen felt that he weighed very little. Which must be to do with the way he was hovering above the floor. He was surprised that his guests didn’t seem to notice (W, GER) DUT De talkshow werd hervat. Hämäläinen voelde zich licht. Wat misschien [‘perhaps’] kwam doordat hij boven de vloer zweefde. Het verbaasde hem dat zijn gasten daar niets van leken te merken. GER Die Sendung wurde fortgesetzt. Hämäläinen fühlte sich leicht. Was damit zusammenhängen mochte [‘could’], dass er über dem Boden schwebte. Er verwunderte ihn, dass seine Gäste es nicht zu bemerken schienen.

(69)

ENG “[…] I don’t think it belongs to the family. They must have rented the place to someone or other”. (VV, IT) DUT “Ik geloof niet dat die van de kinderen is. Ze hebben die plek misschien [‘perhaps’] aan iemand verhuurd”. GER “Ich glaube nicht, dass es den Kindern gehört. Vielleicht [‘perhaps’] haben sie den Platz an jemanden vermietet”.

(70)

ENG The woman always seemed so self-​controlled, so serious. But that must have made things easy for her. (CHL, GER) DUT Zo’n kalme, serieuze vrouw. Misschien [‘perhaps’] heeft juist die uitstraling het haar gemakkelijk gemaakt. GER Diese beherrschte, immer so seriös wirkende Frau. Aber genau diese Ausstrahlung dürfte [‘probably’] ihr den Weg geebnet haben.

The observation that must allows such epistemically weak correspondences can be added to the examples discussed in Lassiter (2014, 2016), in which must is not associated with (strong) certainty either (see examples (3) to (5)). The possibility of such correspondences therefore challenges the position that must is to interpreted as a marker of strong epistemic necessity, as claimed by Gillies and von Fintel 2010.

Inferential must from a Contrastive Perspective

169

5. Conclusion In this paper it has been shown that ENG epistemic-​inferential must does not fully correspond to DUT moeten and GER müssen. We have linked this lacking correspondence mainly to the fact that ENG must spans a wider epistemic range than its Dutch and German counterparts. More in particular, ENG must is compatible with weaker epistemic judgments (probability, possibility, mere assumption) than moeten/​ müssen. Therefore, epistemic-​inferential must can occur in subclauses under non-​factive cognitive verbs (e.g. think, assume, guess), while moeten and müssen tend to avoid such environments. By the same token, must was shown to correspond to relatively weak epistemic markers like DUT waarschijnlijk /​GER wahrscheinlich ‘probably’, DUT vermoedelijk /​GER vermutlich ‘supposedly’ or even DUT mis­ schien /​GER vielleicht ‘perhaps’. This also fits in with the earlier observation (Mortelmans 2012) that ENG must does not only code inferences on the basis of external evidence or world knowledge, but is also used to express speaker-​oriented conjectures that lack a clear evidential basis, which makes them more easily compatible with ‘weak epistemic’ readings. Another contrast to moeten and müssen lies in the fact that must is used to express empathy with the addressee, for which Dutch and German seem to prefer other options. The modals moeten and müssen are less apt to express meanings that take the addressee’s perspective into account or that want to save the addressee’s face. The fact that epistemic-​inferential must behaves differently can be related to the fact that it is more strongly grammaticalized than its Dutch and German cognates. The latter are still prominently geared toward expressing (less grammaticalized) deontic and dynamic modality which typically strongly impacts on the subject (e.g. ENG You must stop/​ DUT Je moet stoppen/​ GER Du musst aufhören). This stark link with the deontic/​dynamic use may have prevented the development of empathic uses for moeten and müssen that we have witnessed for must. By the same token, the strong connection between deontic/​dynamic and epistemic moeten/​müssen might also account for the fact that moeten and müssen have not weakened their epistemic-​inferential meaning to the same extent as must, which is predominantly used with epistemic

170

Tanja Mortelmans

meaning. In fact, weakening of epistemic force –​whereby strong epistemic markers come to express lower degrees of certitude –​has been observed for other strong epistemic markers as well (e.g. surely, I think, no doubt, see Traugott 2011). Finally, what this paper has also shown is how susceptible translators are to subtle language differences that were probably not explicitly described or taught to them. The fact that the translations tend to converge with respect to the phenomena that were the subject of this paper, is quite surprising in itself. At face value, one would expect a simple correspondence between the three verbs. Instead, we get intra-​ and interlinguistic variation which can be accounted for by general principles of grammaticalization.

References Aijmer, Karin 2018. Epistemic must in an English-​Swedish contrastive perspective. In Čermáková Anna /​Mahlberg Michaela (eds) The Corpus Linguistics Discourse. In honour of Wolfgang Teubert. Amsterdam-​Philadelphia: Benjamins, 205–​222. Baumann, Carolin 2017. Bedeutung und Gebrauch der deutschen Modalverben. Lexikalische Einheit als Basis kontextueller Vielheit. Berlin-​New York: de Gruyter. Boye, Kasper 2012. Epistemic Meaning: A crosslinguistic and functional-​cognitive study. Berlin-​New York: de Gruyter. Coates, Jennifer 1983. The Semantics of the Modal Auxiliaries. London: Croom Helm. Close, Joanne /​Aarts, Bas 2010. Current change in the modal system of English: A case study of must, have to and have got to. In Lenker, Ursula /​Huber, Judith /​Mailhammer, Robert (eds) The History of English Verbal and Nominal Constructions. Vol. 1 of English historical linguistics 2008. Selected papers from the 15th International Conference on English Historical Linguistics (ICEHL 15), Munich, 24–​30 August 2008. Amsterdam: Benjamins, 165–​181. Diewald, Gabriele 1999. Die Modalverben im Deutschen. Tübingen: Niemeyer.

Inferential must from a Contrastive Perspective

171

Fintel, Kai von /​Gillies, Anthony S. 2010. Must…stay…strong. Natural Language Semantics 18, 351–​383. Goodhue, Daniel 2017. Must φ is felicitous only if φ is not known. Semantics and Pragmatics 10/​14. https://​semp​rag.org/​index.php/​ sp/​arti​cle/​view/​sp.10.14/​pdf Hogeweg, Lotte 2009. The meaning and interpretation of the Dutch particle wel. Journal of Pragmatics 41, 519–​539 Karttunen, Lauri 1972. Possible and must. In Kimball, John (ed) Syntax and Semantics vol. 1. New York: Academic Press, 1–​20. Kranich, Svenja /​Gast, Volker 2015. Explicitness of epistemic modal marking: Recent changes in British and American English. In Zamorano-​Mansilla, Juan R. /​Maíz, Carmen /​Domínguez, Elena /​Martín de la Rosa, M. Victoria (eds) Thinking Modally: English and Contrastive Studies on Modality. Cambridge: Cambridge Scholars Publishing, 1–​22. Lassiter, Daniel 2014. The weakness of must: In defense of a Mantra. Proceedings of SALT 24, 597–​618. Lassiter, Daniel 2016. Must, knowledge, and (in)directness. Natural Language Semantics 24, 117–​163. Leech, Geoffrey N. 2011. The modals ARE declining: Reply to Neil Millar’s “Modal verbs in TIME: Frequency changes 1923–​2006” [International Journal of Corpus Linguistics 14/​2 (2009), 191–​ 220] International Journal of Corpus Linguistics 16/​4, 547–​564. Letnes, Ole 2010. Zur affektiven Komponente epistemischer müssen-​ Verwendungen. In Katny, Andrzej /​Socka, Anna (eds) Modalität /​Temporalität in kontrastiver und typologischer Sicht. Frankfurt./​Main: Peter Lang, 149–​162. Mihoc, Teodora /​Bhadra, Dita /​Falaus, Anamaria 2019. Epistemic modals, deduction, and factivity: New insights from the epistemic future. Proceedings of SALT 29, 351–​370. Mortelmans, Tanja 2010. Falsche Freunde. Warum sich die Modalverben must, moeten and müssen nicht entsprechen. In Katny, Andrzej /​Socka, Anna (eds) Modalität /​Temporalität in kontrastiver und typologischer Sicht. Frankfurt./​Main: Peter Lang, 133–​148.

172

Tanja Mortelmans

Mortelmans, Tanja 2012. Epistemic must and its cognates in German and Dutch: The subtle differences. Journal of Pragmatics 44, 2150–​2164. Mortelmans, Tanja 2019. Das Modalverb dürfte in epistemischer Verwendung: Ergebnisse einer neuen Korpusstudie. Studia Germanica Gedanensia 41, 113–​126. Nuyts, Jan 2001. Epistemic Modality, Language, and Conceptualization. Amsterdam-​Philadelphia: John Benjamins. Nuyts, Jan /​Byloo, Pieter /​Diepeveen, Janneke 2010. On deontic modality, directivity, and mood: The case of Dutch mogen and moeten. Journal of Pragmatics 42, 16–​34. Squartini, Mario 2008. Lexical vs. grammatical modality in French and Italian. Linguistics 46/​5, 917–​947. Squartini, Mario 2016. Interactions between modality and other semantic categories. In Nuyts, Jan /​van der Auwera, Johan (eds) The Oxford Handbook of Modality and Mood. Oxford: Oxford University Press, 50–​67. Traugott, Elisabeth Closs 2003. From subjectification to intersubjectification. In Hickey, Raymond (ed) Motives for Language Change. Cambridge: Cambridge University Press, 124–​139. Traugott, Elisabeth Closs 2011. Modality from a Historical Perspective. Language and Linguistics Compass 5/​6: 381–​396. Ulvestad, Bjarne 1984. Die epistemischen Modalverben werden und müssen in pragmalinguistischer Sicht. In Stickel, Gerhard (ed) Pragmatik in der Grammatik. Jahrbuch 1983 des Instituts für deutsche Sprache. Düsseldorf: Schwann, 262–​294. Van der Auwera, Johan /​Plungian, Vladimir 1998. Modality’s semantic map. Linguistic Typology 2, 79–​124. Wiemer, Björn 2018. Evidentials and epistemic modality. In Aikhenvald, Alexandra (ed) The Oxford Handbook of Modality. Oxford: Oxford University Press, 85–​108.

Aoife Ahern, José Amenós-​Pons and Pedro Guijarro-​Fuentes

Conjectural Future in French and in Spanish: An L2 Acquisition Perspective

Abstract The Romance future tense (FUT) can encode evidential meaning, related to conjecture. However, across the different Romance languages, diverse semantic and pragmatic factors limit the available interpretations of FUT. We examine the lexical aspect restrictions that are involved in such interpretations in French and in Spanish, and some of their consequences for the acquisition of Spanish by French-​speaking learners. The commonalities and contrasts on how FUT can be interpreted across the two languages may provide certain advantages but also pose challenges, since a full command of this tense would entail feature reassembly. An empirical study is presented on the acquisition of conjectural future in L2 Spanish by L1 French speakers (n=​33) at upper-​intermediate and advanced levels, and contrasted with data from a control group of native European Spanish speakers (n=​35). The data suggest that the L2 learners, even at advanced level, continue to differ significantly from native speakers in their interpretation of conjectural future under the range of semantic and pragmatic conditions tested. Keywords: Future tense, Conjectural interpretations, Aspectual restrictions, L1 French, L2 Spanish

1. Synthetic Future Tense in Romance: an Overview In modern Romance languages (except for Romanian and some Italian vernaculars), a synthetic, inflectional future form (henceforth, FUT)

174

Aoife Ahern et al.

exists, having developed from a periphrastic source: the Latin obligative construction ‘Infinitive +​habere’. Futurity can also be expressed in the same languages by other means, in addition to FUT, such as verbs of movement (go and come), volition, markers of progressive aspect or certain uses of simple present. Full grammaticalization of these alternative devices in the respective languages, however, is uncommon (Dahl 2000). Since the expression of futurity in Romance is etymologically related to modality, it is not surprising that, in most Romance languages, the FUT forms may receive not only interpretations related to chronology but also conjectural interpretations, similar to those of epistemic modal expressions, as in (1) and (2): (1)

(Echador de cartas) Te casarás y tendrás muchos hijos. [(Fortune teller) You’ll marry and you’ll have lots of children]

(2)

No ha telefoneado. Estará ocupado. [He has not phoned. He must be (FUT) busy]

Interpreting chronology and conjecture often leads to the identification of additional nuances such as (on the chronological side) directiveness and volition, as well as (on the conjectural side) concession and mirativity, as in the Spanish examples (3) to (6): (3)

No se lo dirás a nadie; es una orden. (directiveness) [You will not tell anyone; that’s an order]

(4)

Te lo traeré mañana, no te preocupes. (volition) [I’ll bring it to you tomorrow; don’t worry about it]

(5)

Será muy guapo, pero es muy antipático. (concession) [He may be (FUT) very handsome, but he’s very unpleasant]

(6)

¡Será tonto…! (mirativity) [He’s (FUT) such a chumb!]

However, the restrictions that apply to such readings differ across the Romance languages. The contrast is particularly dramatic in the

Conjectural Future in L1 French / L2 Spanish

175

two languages that will be studied here, namely Spanish and French. In the present article, we will examine the nature of this contrast, and how it affects the L2 acquisition of the Spanish FUT by L1 French speakers. If L1 transfer is assumed, its advantages and disadvantages, as experienced by the L2 learners, will reflect the crosslinguistic similarities and differences in this shared linguistic resource, whose semantics is frequently presumed to be cross-​linguistically identical. In Spanish, conjectural interpretations correspond to a very high percentage of the simple FUT usage (Escandell-​Vidal 2010, 2014), while the mere existence of these interpretations in present-​day spontaneous French is a debated issue (Le Querler 1996; Tasmowski and Dendale 1998; Rocci 2000; De Saussure and Morency 2012). However, despite the frequency of conjectural simple FUT in Spanish, linguistic restrictions also apply in this language; they are mainly related to Aktionsart (Soto 2008). For conjectural interpreta­ tions of the simple FUT, the verb usually needs to have [-​telic] and [-​dynamic] features (as in (7) and (8)). Therefore, states are naturally compatible with such readings. On the other hand, [+​dynamic] verbs in the simple FUT can only be interpreted as conjectural if their location in the future is contextually ruled out. This generally implies that the events denoted are understood as habitual (10): (7)

(Llaman a la puerta) Será Juan. [-​telic] [-​dynamic] [(They are knocking at the door) It will be Juan.]

(8)

Será muy listo, pero no lo parece. [-​telic] [-​dynamic] [He may be (S-​FUT) very smart, but does not seem to be.]

(9)

?? No coge el teléfono. Preparará la cena. [+​telic] [+​dynamic] [S/​he’s not answering the phone. S/​he must be making (S-​FUT) dinner.]

(10)

-​¿A qué se dedica Luis? +​No sé, trabajará en cosas de informática. [-​telic] [+​dynamic] [+​habitual] [-​What does Luis do? +​I don’t know, he (probably) works (S-​FUT) in computing.]

176

Aoife Ahern et al.

In Spanish, the aspectual restrictions mentioned above can be overridden by adding a [+​progressive] feature to the predicate,1 as in (11) and (12): (11) No coge el teléfono. Estará preparando la cena. [S/​he’s not answering the phone. S/​he must be making (PROG S-​FUT) dinner.] (12) Estará preparando la cena, pero no le he visto en la cocina. [He may be making (simple FUT +​progressive) dinner, but I haven’t seen him in the kitchen.]

As regards conjectural interpretations of the compound FUT, they obtain in both Spanish and French without any lexical aspect restrictions. The compound FUT is unlike the simple form in that it involves the representation of a completed event, or its resulting state, as being previous to a reference point. This facilitates non-​future readings regardless of the predicate type, as in utterances (13) and (14), in which conjectural interpretations in Spanish and French are found with achievements: (13) La puerta no ha sido forzada. El ladrón habrá entrado por la ventana. [The door has not been forced. The burglar must have entered (compound FUT) through the window.] (14) Il est midi. Max aura (déjà) atteint le sommet. [It’s 12 o´clock. Max must (already) have reached (compound FUT) the summit.]

A further difference between French and Spanish is found in concessive environments, like (8) and (12) above. Concessive FUT constructions are common in Spanish, with both the simple and compound

1

Present-​day French lacks a fully grammaticalized progressive device. The con­ struction être en train de +​infinitive can be used to express progressivity in the future, but its frequency is much lower than the Spanish be +​gerund progressive construction (Bertinetto 2000). This may explain why, in French, non-​stative predicates tend to be incompatible with conjectural readings.

Conjectural Future in L1 French / L2 Spanish

177

FUT; but in French, neither the simple FUT nor its compound counterpart can be used to express conjecture: (15)

*Il sera intelligent, mais il ne le semble pas. Il a beau être intelligent, mais il ne le semble pas. [He may be very smart, but does not seem to be.]

(16)

*Il aura préparé le dîner, mais je ne l’ai pas vu dans la cuisine. Il a peut-​être préparé le dîner, mais je ne l’ai pas vu dans la cuisine. [He may have made dinner, but I haven’t seen him in the kitchen.]

A thorough description of the different usages of FUT in the languages of our study is beyond the scope of this paper. However, the overview provided in the previous paragraphs clearly shows that despite its common origin, the FUT tense has developed opposite tendencies in Spanish and French, particularly in terms of its potential for expressing evidential meaning. Therefore, it should not be taken for granted that the semantics of FUT is still the same in both languages, and this is not a trivial matter, because any hypothesis on the learning task of L1 French speakers acquiring L2 Spanish will necessarily rely on specific assumptions on the (semantic and/​or pragmatic) nature of the cross-​linguistic contrast. Furthermore, both languages lack specialized grammatical devices that encode evidentiality, yet share the characteristic, along with other Romance languages, of conveying evidential meaning through certain ‘evidentiality strategies’ (as per Squartini 2008); little is known so far of how these kinds of similari­ ties and differences may affect the process of acquiring such structures and successfully associating them with the discourse conditions that restrict them. The remainder of this paper will be structured as follows: in section 2, different proposals on the semantics of the FUT in Spanish and French will be considered. In section 3, we provide an overview of how the effects of L1 transfer have been described in the literature on L2 acquisition, followed by an outline of the theoretical background for the present study. In section 4, we report on the implementation and results of an empirical study of L2 Spanish FUT acquisition. Discussion and conclusions will be presented in section 5.

178

Aoife Ahern et al.

2. The Semantics of FUT in Spanish and French In French, the conjectural interpretations of the FUT have generally been described as cases of contextual enrichment of a core chronological meaning: they involve the future verification of an assertion about the present –​or, with compound FUT, about the past (Wilmet 1996; De Saussure and Morency 2012). This claim has also been made in studies on the Spanish FUT (e.g. Rojo and Veiga 1999; Rodríguez Rosique 2015), but it is far from consensual, and difficult to reconcile with facts such as that synchronically the expression of conjecture is, at least in some American varieties of Spanish, the most frequent use of the FUT (Escandell-​Vidal 2010, 2014). In this vein, concerning French, De Saussure and Morency (2012) describe the pragmatic requirements that map onto the aspectual asymmetries described in the previous section: conjectural uses of FUT are only acceptable if they represent an event that is verifiable in the future by an allocentric third party. The connection of present and future time is straightforward with states, since they represent situations that are expected to continue for a relevant time; conversely, achievements lack such a property, so they are generally incompatible with conjectural readings. As to the [-​dynamicity] requirement, the oddity of activities and accomplishments in conjectural, non-​progressive simple FUT is due to the fact that the continuity of the event is not semantically entailed, so it must be pragmatically inferred. This implies a supplementary processing effort, which the hearer will only undertake if additional cognitive effects are warranted (De Saussure and Morency 2012: 214). The future-​verification path, supplemented by contextual enrichment processes, is an intuitively appealing idea, properly accounting for the severe restrictions of the conjectural FUT in French. Still, the fact that the different Romance languages exhibit dissimilar restrictions remains unexplained. Comparing French and Spanish, Azzopardi (2017) concludes that the cross-​linguistic dissimilarities are not due to the semantics of the FUT itself, but to the fact that the two languages have developed alternative ways of expressing modality. In French, the FUT can only express a conjecture related to abductive reasoning; i.e. a conceivable, but

Conjectural Future in L1 French / L2 Spanish

179

not necessarily likely, hypothesis. Contrastingly, the use of the periphrasis devoir ‘must’ +​infinitive is compulsory if deductive reasoning is involved. This constraint does not exist in Spanish; therefore, some usages of the conjectural FUT in Spanish are not acceptable in French (even in the compound form), if the context suggests a high degree of likelihood:2 (17)

No habrá oído nada, señor. Esta mañana la señora padecía jaqueca y no quería levantarse de la cama. (Eduardo Alonso, Flor de Jacarandá)

(18)

Elle n’a pas dû entendre, Monsieur. Ce matin Madame avait la migraine et elle ne voulait pas se lever du lit. [She won’t have heard anything, sir. This morning madam had a migraine and didn’t want to get up out of bed.]

(19)

-​ Ahora estará en la biblioteca, aventuró la azafata. -​¿En la biblioteca? –​Madame lee muchos libros. Es su pasatiempo favorito (Eduardo Alonso, Flor de Jacarandá)

(20)

-​En ce moment elle doit être dans la bibliothèque -​hasarda la femme de chambre. –​Dans la bibliothèque ? –​Madame lit beaucoup de livres. C’est son passe-​temps favori. [Now she’ll be in the library, the hostess hazarded.] [In the library?] [Madam reads lots of books, it’s her favorite pastime.]

Azzopardi satisfactorily accounts for some of the contextual dissimilarities between French and Spanish. Yet, the widespread availability of non-​future uses of the FUT tense in different languages (and not only in Spanish) remains unexplained. Considering the overall Romance variation, Escandell-​Vidal (2021: 18) chooses a different path. She describes the semantics of the Romance FUT in terms of features; there are certain core semantic 2

Both pairs of sentences, (15–​16) and (17–​18), show the French translation of utterances from Alonso’s text, also corresponding to the English translations provided by Azzopardi.

180

Aoife Ahern et al.

features common to all Romance varieties, and others that differ. From her point of view, the core meaning of the Romance FUT ‘encodes an instruction [for the hearer] to build the eventuality described by the sentence radical, or prejacent proposition p, as the representation of a situation (ES) that cannot be accessed from an information acquisition situation (IS), in which the discourse situation (DS) is included’. The core semantic features do not specifically relate to time, nor to conjecture, but to the information source: the situation described is (at the IS) only available as an inference of the speaker. This implies that the FUT has an inherent evidential meaning.3 Despite the existence of common Romance features, the ultimate materializations of the FUT tense in the different languages will vary depending on the values that other cross-​linguistically shared features have taken in each of them. These features and their variation are represented below in Table 1: Table 1.  The semantics of the simple FUT in Romance (Escandell-​Vidal 2021). IS ǀ ES DS ⊂ IS

CORE MEANING PARAMETERS

CATALAN

FRENCH

ITALIAN

SPANISH

PORTUGUESE

ES [± Factual] IS [± Deictic] IS ǀ ES [± Forward]

-​

-​

-​

-​

-​

-​

+​

+​

+​

+​

+​

+​

-​

-​

-​

Thus, following Escandell-​Vidal’s proposal, the Romance future is always non-​factual. In all languages but Catalan it is also [+​deictic]; this means that (in these languages) the IS is related to the speaker’s perceptual field. The Catalan FUT is the only one in Romance that is independent from it; as a result, this language disallows conjectural 3

The idea that the Spanish FUT is intrinsically related to evidentiality has inde­ pendently been explored by other linguists, such as Squartini (2001, 2004), Matte Bon (2007), Laca and Falaus (2014), Laca (2016) and García-​Negroni (2016).

Conjectural Future in L1 French / L2 Spanish

181

future (at least in the natural, spontenous use of most varieties), as opposed to French, Italian, Spanish and Portuguese. However, the French FUT imposes a [+​Forward] relation between IS and ES: thus, all eventualities in the French FUT call for a future verification. This is not required in Italian, Spanish and Portuguese and, as a result, conjectural uses of the FUT are much more widespread. Escandell-​Vidal’s hypotheses on the semantics of the FUT underlie our own study because they have a clear advantage over alternative proposals since they provide explicit, falsifiable linguistic criteria for the commonalities and distinctions of the uses of FUT in Romance. They also offer a basis for conceptualizing the learning task of the L2 learners of our study: a speaker of L1 French, in order to fully acquire the Spanish FUT, must reassemble the [± Forward] feature. This claim has a direct connection with current L2 acquisition theory and will be addressed in the following section.

3. Language Transfer and Feature Re-​assembly 3.1. Language Transfer in L2 Acquisition The matter of language transfer during L2 acquisition has been an important research topic since its conceptualization. One of the founders of the notion of language transfer was Weinreich, who considered this term for ‘instances of language deviation from the norms of either language which occur in the speech of bilinguals as a result of their familiarity with more than one language’ (Weinreich 1953: 1). Some­ what more recently, Odlin (1989: 27) deemed that ‘transfer was the influence resulting from similarities and differences between the target language and any other language that has been previously (and perhaps imperfectly) acquired’. Odlin established a distinction between positive (or facilitative) transfer and negative transfer phenomena (e.g. underproduction or overproduction of a particular structure, production errors such as substitutions, calques and alterations of a target language item, misinterpretations during comprehension). Other researchers,

182

Aoife Ahern et al.

however, including Sharwood-​Smith and Kellerman (1986: 1), prefer to limit the notion of transfer to referring just to ‘those processes that lead to incorporation of elements from one language to another’. Reflecting on some of the research carried out within an innatist view of language acquisition, the L1 system could influence the second language acquisition (SLA) process in different significant ways; yet, not all researchers within the generative approach agree on the importance of the native language effect on SLA (Gabriele, 2009). Since the 1990s different hypotheses have flourished in SLA. One such hypothesis is the Full Access/​Full Transfer hypothesis (e.g. Schwartz and Sprouse, 2000; White, 2003), under which it is assumed that features (interpretable and uninterpretable alike) that do not form part of the L1 can be attained by adult L2 learners. This account has long hypothesized that L1 transfer is a potential source of L1/​L2 developmental sequence and ultimate attainment differences, asserting that L1 transfer significantly (and variably –​depending on the L1/​L2 pairing) influences the L2 learning task. Under the postulation that the learning task for L2 adults is not essentially different from that of the L1 child, the Interpretability Hypothesis (henceforth, IH) (Hawkins and Hattori 2006; Tsimpli and Dimitrakopoulou 2007) asserts that parameter (re)setting is only limited to L1 interpretable feature values, and not applicable to uninterpretable features. That is, only those features instantiated within the L1 and, at most, new interpretable features stay available in SLA. Uninterpretable features, which have syntactic effects, are contended to no longer be available to adults due to critical period limitations. This would explain an L2 adult’s inability to reset parameters, at least when they are contingent on the acquisition of new uninterpretable features. The IH contends that the underlying syntax of L2 grammars is destined to remain like that of the L1 grammar with localized (or surface) adjustments. Nonetheless, other accounts have purported that L2 variability does not point to deficits with uninterpretable features, as claimed by the IH, but emerges from interface vulnerabilities. For instance, interpretable features relevant to the syntax–​semantics or the syntax–​ discourse interfaces are problematic. These kinds of proposals seem to take into account that L1 transfer alone cannot account for all L2 variability (see Sorace, 2011, 2012 for a review of interface vulnerability

Conjectural Future in L1 French / L2 Spanish

183

in L2 theorizing). In standpoints which consider L1 transfer and/​or inaccessibility to features as the primary source of L2 variability, it is predicted that L2 learners whose L1 instantiates similar features for a given property should not display significant variation for those features in the L2. It has already been shown that L2 variability is not exclusively selective on a continuum of what can be transferred from the L1. That is to say, L2 learners whose L1s should supply them with features that are the same in both languages also reveal different degrees of variability in applying properties connected with those very same features; mainly when such features involve interfaces. On the contrary, if interface properties are simply more difficult for all L2 learners, then this fact can account for much L2 variability, even when it is seemingly unexpected given L1 transfer and similarities in the two languages. More recently, the mapping challenges that L2 learners with different L1 backgrounds may face has contributed to an intense debate. In particular, the acquisition of syntactic features has allowed SLA researchers to investigate whether L2 learners are able to recognize the different behavior and nature of these features, and to explore whether the difficulties in attaining them are due to any underlying syntactic deficit. In that respect, the Feature Reassembly Hypothesis (henceforth, FRH) (Lardiere 2008, 2009) argued that L2 learners can acquire subtle features and that learners reassemble the sets of lexical features of the native language into feature bundles appropriate to the L2. 3.2. Transfer in Closely-​related Languages: L1 French /​L2 Spanish Taking stock of the level of interaction between the L1 and L2 language systems, it is clear that L1 transfer is a possible source of L2 divergences, and that L1 transfer significantly (and variably, depending on the L1/​L2 pairing) affects the L2 acquisition task. Nonetheless, despite all the progress made so far, one of the many remaining questions related to language transfer is the nature of the interaction that it triggers when it takes place between two similar languages. One might hypothesize that, during acquisition processes involving an L2 that is typologically related to the L1, learners may form a kind of ‘hybrid

184

Aoife Ahern et al.

interlanguage’, which would consist of features from the two language systems, or alternatively learners may even go a step further by creating a system which does not exist in either their L1 or L2. In the previous section, following Escandell-​Vidal (2021), it has been argued that the semantics of the FUT in Spanish and in French is partially different; those differences have a direct effect on the availability of conjecture-​related interpretations in each language. Therefore, it could be envisaged that L1 transfer will have a negative effect on L2 acquisition. However, it is also possible that language similarities may not hamper the acquisition process, if some but not all of the L1 features are transferred. A glance at the scarce literature on transfer across similar languages, with particular reference to the acquisition of French and Spanish considering different grammar features (e.g. adjectival placement, number, object clitics, null objects), different learner populations and an array of elicitation methods (Guijarro-​Fuentes 2014, 2015; Bruhn de Garavito 2010; Grüter and Crago 2010, to name just a few), shows mixed findings depending on the features in question. For instance, L2 Spanish learners of French acquiring object clitics and null objects have little or no difficulty with such features (Grüter and Crago 2010), suggesting a positive role for transfer; however, its advantages hinge on the linguistic task (production versus comprehension). Bruhn de Gar­ avito (2010), on the contrary, found variability in the production of the Spanish plural feature by L2 French speakers, contrary to predictions, since both French and Spanish possess strong number features; by contrast, the L2 speakers did not have problems with the acquisition of noun/​adjective word order. She concluded that negative transfer at one level (e.g. prosodic) can affect other levels of the grammar (e.g. syntax). In short, findings from previous studies seem to suggest that L1 influence can lead to a certain degree of facilitation for certain individual features. Closer to the purposes of our present study, recent investigations (Ahern, Amenós-​Pons and Guijarro-​Fuentes 2020) on the acquisition of a range of Spanish FUT interpretations by French-​speaking learners indicated that despite certain advantages (i.e. positive transfer) in the use and interpretation of chronological FUT, learners produced the tense significantly less than native speakers for expressing conjectural

Conjectural Future in L1 French / L2 Spanish

185

and concessive meaning. In the study, involving upper intermediate and advanced-​level learners of Spanish (29 with CEFR B2 level of proficiency and 19 with C1) in formal learning environments, the results of three linguistic tasks were compared to those of a control group of 29 native speakers of European Spanish. Despite the difference in the production of conjectural and concessive FUT uses, the learners’ understanding of all these uses of FUT was shown to be similar to that of L1 speakers. However, in the study it was unclear whether the L2 learners were actually aware of the linguistic restrictions operating in Spanish, which are different from those of French (as described in section 1 above). The present study intends to shed light on the specific issue of the acquisition (by L1 French speakers) of the linguistic (mainly aspectual) restrictions on the expression and interpretation of evidentiality by means of the conjectural and concessive Spanish FUT. In this paper, we describe a comparative study between two groups of adult L1 French speakers at two different proficiency levels of L2 Spanish (CEFR B2 and C1), albeit similar in terms of the learning experiences. The following questions are therefore explored here: 1. 2. 3.

Will second language learners transfer all L1 interpretable features into their L2 target language, which in the present case could hinder the learning process? Will negative transfer and developmental errors evolve (increase or decrease) comparing the two proficiency levels considered here? Will the learners eventually acquire those uses of the Spanish FUT, limited by different restrictions in their L1 (i.e. will they eventually be able to reassemble the concerned features)?

4. The Interaction between Lexical Aspect and Spanish Future Tense Interpretations in L1 and L2 Spanish Considering the characteristics of the FUT in French and Spanish, and the hypotheses mentioned above in relation to the semantics of FUT

186

Aoife Ahern et al.

and to the acquisition process by which French L1 speakers develop full proficiency in Spanish, we now report on a set of empirical tasks that were designed and applied as part of the current study, and present the results obtained. 4.1. Empirical Tasks and Methodology With aims that are subsumed under the more general research questions brought up above at the end of Section 3, our study on the restrictions on interpreting conjecture and concession readings of Spanish FUT tense consisted of a set of tasks designed to detect evidence of sensitivity to these restrictions in both native speakers of Spanish and L1 French learners of Spanish. This study follows up on results of previous research (Ahern, Amenós-​Pons and Guijarro-​Fuentes 2020) in which data on the acquisition, by French L1 learners, of conjectural and concessive interpretations of Spanish FUT were presented. The learner groups showed progress, as their general proficiency in L2 Spanish increased, in the understanding of conjectural FUT uses, but not in concessive interpretations. However, this and other previous studies lacked in-​depth consideration of the degree to which the lexical aspect of the verbs included affected the results of the linguistic interpretation tasks. 4.2. Participants The participants in this set of tasks included, therefore, L1 Spanish speakers, and L1 French speakers who were students of Spanish at the Instituto Cervantes in Paris, Lyon or Toulouse. A total of 36 native speakers (central variety of Peninsular European Spanish), and thirty-​three L2 learners participated by way of the Internet, firstly completing an ethnolinguistic questionnaire and then the experimental linguistic tasks. The learner groups included ten participants who had (CEFR) B2 level of general proficiency in Spanish, and twenty-​three who were at C1 level. The native speaker group included forty-​three adults (twenty-​ nine women and fourteen men), all born in Spain, all of whom also

Conjectural Future in L1 French / L2 Spanish

187

had knowledge of other languages, mainly English, but none of whom considered themselves bilingual: twelve participants identified their English knowledge as being at intermediate level; seven of them, upper intermediate (B2); ten with advanced English. Of these participants nineteen indicated knowledge of French, of which the following mentioned their level of proficiency: one beginner, seven with upper elementary level (A2), five at intermediate level, one upper intermediate, four advanced. There were also nine of these participants who mentioned, in addition to English, knowledge of some other language besides French (Chinese, Greek, German, Flemish, Norwegian or Portuguese). The French-​speaking group of L2 adult Spanish learners was made up of nineteen women and fourteen men. One of the participants was a monolingual French speaker, whereas the other thirty-​two all indicated that they had some knowledge of English, many describing their command of this language as ‘school English’; while only four stated that they had advanced levels of English. In terms of the order in which they had learnt the languages, only two participants had learnt Spanish as their first foreign language; two had learnt German as their first foreign language and all the others had learnt English as their first foreign language, before learning Spanish. With regard to immersion experiences, only one of these Spanish learners had lived in a Hispanic country (in Barcelona, for seven years), while another learner had studied in Bolivia for several short periods. 4.3. Task and Items Both linguistic tasks consisted of grammaticality judgement tests. The first task included nineteen focus items and five distractors, and the second task included eighteen focus items and five distractors. The items were presented to the participants in randomized order. The items firstly presented a stimulus including a contextualized utterance in which FUT tense appeared, with the focus verb form highlighted by means of block capital letters. Below each stimulus, three judgement options were shown, of which the participants chose one and clicked on a check-​box to indicate their choice (see example items in Appendix I). The options to choose from were the following: (a) The verb form in

188

Aoife Ahern et al.

block capitals is right /​(b) It sounds strange; I’m not sure that this form is right /​(c) The verb form is wrong. The first of these two tasks (hereon, Task 1; item conditions shown in Table 2) focused on the relative acceptability of conjectural and concessive FUT interpretations in relation to the lexical aspect –​ Aktionsart –​of verbs. This task presented a linguistic context (a brief text) followed by the stimulus and the three judgement options just mentioned. The second task (Task 2) presented an “external” stimulus, specifically, an illustration or a simulated newspaper headline and strapline, shown together with a written utterance in which a FUT form appeared in block capitals, and the same judgement options as those in the previous task, shown below the stimulus. While in Task 1 all the FUT forms were simple, in Task 2, half of the FUT forms presented were compound (e.g. habrá hecho), and the FUT items did not include any concessive uses, but only conjectural ones, as the task consisted in verbalizing possible interpretations of what the situation shown in the external stimulus was or how it had arisen. Thus, Task 1 focused on the effect of the lexical aspect on the acceptability of conjectural and concessive interpretations of the simple FUT (which are felicitous only with atelics, particularly with states), while Task 2 focused on the different acceptability of the simple and the compound FUT for conjectural interpretations (for the compound FUT, conjectural interpretations are felicitous with all predicate types). The item conditions of both tasks are displayed in ­tables 2 and 3. In Appendix I, examples are provided for the items corresponding to the conditions shown in these tables. Table 2.  Task 1 item conditions. Item numbers

Interpretation of FUT

Lexical aspect of verb in FUT

conjecture

concession

state

1 and 2 3 –​5 6 –​8 9 –​11 12 –​15 16 –​19

√ √

√ √

√ √ √ √

activity

telic

√ √ √ √

189

Conjectural Future in L1 French / L2 Spanish Table 3.  Task 2 item conditions. Item numbers

Tense Simple FUT

25, 33, 41 26, 34, 42 27, 35, 43 28, 36. 44 29, 37, 45 30, 38, 46

Compound FUT

√ √ √ √ √ √

Interpretation of FUT

Lexical aspect of verb in FUT

conjecture

state

√ √ √ √ √ √

concession

activity

telic

√ √ √ √ √ √

Considering the hypothesis of an evidentiality-​based semantics of Spanish FUT (Escandell-​Vidal 2010, 2014), and the predictions on the value-​mapping contrast between French and Spanish shown in Section 2, our hypotheses for each item set were as follows: 1.

2.

3.

4.

Stative verbs in FUT: Native speakers will accept these uses with ease. If L1 transfer is operating, L2 learners will also tend to accept the conjectural interpretations but hesitate to accept items with concessive readings due to its inexistence as a possible interpretation of the corresponding L1 (French) forms. However, considering that, in everyday French, conjectural uses of FUT are extremely infrequent, L1 transfer might also limit the acceptability of the conjectural items, even with states. Activities in FUT: Native speakers will tend to only accept these items in cases where a habitual reading is made accessible in the discourse context (as in items 6 to 9). Due to L1 transfer, L2 learners will tend to reject all the items with this pattern. Accomplishments and achievements in FUT: Both the native speakers and the learners are expected to reject simple FUT in sentential contexts consistent with hypothetical and concessive interpretations, because of the telic nature of these lexical aspect classes. Compound FUT forms: In the items with compound FUT, it was hypothesized that both native speaker and learner groups would

190

Aoife Ahern et al.

accept telic and atelic verb forms, since conjectural use of the compound FUT is available in Spanish and in French. Thus, in Task 1, items 1 and 2 are expected to be widely accepted in all groups, although more variability is predicted in the L2 group; items 3 to 5 and 9 to 11 will only be clearly accepted by the L1 group; items 6 to 8 and 12 to 19 should be rejected by all groups. Regarding Task 2, for the items involving simple FUT forms, those with stative predicates (items number 25, 33 and 41) should be accepted (to dissimilar degrees) by natives and non-​natives; whereas all the items with verbs expressing activities, achievements or accomplishments (items 26, 27, 34, 35, 42 and 43) should be rejected. Finally, the items containing compound FUT forms (items 28 to 30, 36 to 38 and 44 to 46) should be accepted by all groups, with less variability than in the simple FUT items. 4.4. Results The results of Task 1 are displayed in ­figures 1 and 2 below; Task 2 results are shown in Figure 3: 100 90 80 70 60 50 40 30 20 10 0 a

b

c

a

B2/11 Conjecture States

b

c

C1/23 Conjecture Activities

Figure 1.  Conjecture items (Task 1).

a

b CONTROL/36

Conjecture Telics

c

191

Conjectural Future in L1 French / L2 Spanish 100 90 80 70 60 50 40 30 20 10 0 a

b

c

a

b

B2/11

c

a

b

C1/23 Concession States

c

CONTROL/36

Concession Telics

Figure 2.  Concession items (Task 1).

100 90 80 70 60 50 40 30 20 10 0 a

b

c

B2/11

a

b C1/23

c

a

b CONTROL/36

FS States

FS Activities

FC States

FC Activities FC Telics

Figure 3.  Task 2 results.

FS Telics

c

192

Aoife Ahern et al.

The statistical analyses of the results included Shapiro-​Wilk tests, which rejected a normal distribution of the overall results across groups. Therefore, non-​parametric tests were employed. The performances across the three groups were compared using Kruskal-​Wallis tests; two-​group comparisons were also done by means of Mann-​ Whitney tests. As shown in ­figures 1 and 3, in the items where the discourse context determined conjectural interpretations of the simple FUT, the L1 Spanish speakers’ responses were clearly conditioned by the verbs’ lexical aspect: in task 1, 88% of the participants accepted the stative items; in Task 2, 54.6% of the stative items were accepted. Contrastingly, only 34.2% of the participants accepted the items involving activities in Task 1, and 17.5% in task 2. Finally, conjectural interpretations of the simple FUT with telic predicates were clearly rejected (only 9.9% in Task 1 and 9.2% in Task 2 considered them correct). The differences between these three item groups (states, activities and telics) were significant (p < .05). In answering the same groups of items, the L1 French speakers were also conditioned by lexical aspect, but to a significantly lesser degree than the L1 Spanish participants: in Task 1, 70% (at B2) and 58.6% (at C1) L1 French speakers considered the items involving states as correct. The percentage was very close for activities: 56.6% (at B2) and 55% (at C1). In Task 2, 36.6% (at B2) and 33.3% (at C1) L1 French participants accepted conjectural constructions with states, and 26.6% (at B2) and 33.3% (at C1) accepted those constructions with activities. For atelic predicates, all the differences between the L1 and the L2 groups were significant (p < .05). In contrast, the L1-​L2 differences for the telic predicates in simple FUT were not significant (p > .05). The L1 French speakers rejected conjectural constructions with telics as clearly as the members of the L1 Spanish group (in task 1, telic items were accepted only by 15.7% of the B2 participants and by 14.9% of the C1 participants; in task 2, the percentages were 20% at B2 and 10% at C1). The responses of the two L2 groups (B2 vs C1) showed no significant difference for any item group. As for the concessive interpretations of the simple FUT, the response patterns shown by L1 and the L2 participants differ significantly (p > .05). The native Spanish speakers accepted this interpretation

Conjectural Future in L1 French / L2 Spanish

193

at a rate of 86% with states; in contrast, only 48.3% (at B2) and 44.9% (at C1) L1 French speakers considered the concessive items with states correct. On the other hand, the native and non-​native reactions to telic predicates were analogous, without any statistically significant differences: only 7.5% (L1 Speakers), 10% (L1 French speakers at B2) and 21% (L1 French speakers at C1) judged these items as correct. Finally, no clear aspectual bias nor significant differences were found between the groups in their acceptance of the compound FUT in items expressing concession. In the L1 Spanish group, the percentages of acceptance are: 75% for states, 81.4% for activities and 91.6% for telic verbs. The percentages in the L1 French groups are: 80% (at B2) and 69% (at C1) for states; 86% (at B2) and 76.8% (at C1) for activities; 83% (at B2) and 76.8% (at C1).

5. Discussion and Conclusions Our results provide support to most of the hypotheses introduced in Section 4.1. Natives were sensitive to lexical aspect effects; they clearly accepted all the items involving states, with no differences between conjectural and concessive constructions (hypothesis 1). Non-​natives also favored states, but to a significant lesser degree, and discriminating between conjectural and concessive items. Therefore, the infrequency of the conjectural FUT in their L1 and the non-​existence of concessive constructions using FUT have a clear effect in the answer pattern. The answers given to activities (hypothesis 2) create an interesting contrast between natives and non-​natives. As expected, L1 Spanish speakers effectively discriminate between the naturalness of conjectural interpretations of states (widely accepted) from those of activities (which are often rejected). This difference is not found in the L1 French groups: contrary to our expectations, the percentage of acceptance of states and activities is rather close. This strongly suggests that feature reassembly has not really taken place; the L2 learners are not sensitive to fine-​grained aspectual distinctions.

194

Aoife Ahern et al.

Nevertheless, the overall picture is nuanced if we consider the contrast between atelics vs telics (hypothesis 3). As expected, when telic verbs in the simple FUT are involved, natives and non-​natives perform similarly; all the groups tendentially reject the expression of conjecture and concession with those verbs. This suggests that the [± telic] feature that distinguishes telic vs. atelic predicates is more prominent (for the non-​natives) than the [± dynamic] feature that separates states and activities. It would be tempting to conclude that feature reassembly starts with the [± telic] feature, before extending to the [± dynamic] feature. However, without data on proficiency levels below B2 and over C1, no support to that hypothesis can be provided by the present study. Finally, the fact that L1 French speakers clearly differentiate simple vs compound FUT (hypothesis 4) provides new evidence on the L1 transfer effect: the expression of conjecture with the compound FUT is possible in French; therefore, the L1 French participants tend to accept it, almost as much as the L1 Spanish speakers. Taken together, our results indicate that L1 transfer is the main force guiding the L2 judgements of our non-​native participants; no evidence of any awareness of the [-​Forward] feature (which facilitates the availability of conjectural and concessive use of the simple FUT in Spanish) was found. Still, the fact that telic and atelic verbs create different answer patterns could tentatively suggest that feature reassembly will eventually take place if more input is available or, perhaps, if more attention is given in the classroom to the cross-​linguistic differences in the use of FUT in Spanish and French. Turning now to the three research questions we introduced at the end of Section 3 (repeated here for convenience), our data deliver the following answers: 1.

2.

Will second language learners transfer all L1 interpretable features into their L2 target language? Full transfer of L1 features has been clearly detected; our L2 participants tend to apply the restrictions operating in their native language. This has a hindering effect regarding the uses of the simple FUT considered here, but leads to a facilitative effect in the use of the compound FUT. Will negative transfer and developmental errors evolve? No significant differences were found between the two L2 proficiency

Conjectural Future in L1 French / L2 Spanish

3.

195

levels considered here. This suggests that feature reassembly is a long-​lasting source of difficulty. Will the learners eventually acquire the uses of the Spanish FUT that have different restrictions in their L1? No definite answer can be given in the present state of our research. Still, the fact that the L2 participants showed sensitivity to the [± telic] feature suggests some positive evolution (in the learning trajectory that led up to the levels considered here) and opens the possibility of subsequent progress.

In the literature on tense and evidentiality (see references in Section 2), conjectural and concessive uses of the FUT are often described as a way to specify an information source-​type; following Escandell-​Vidal (2010, 2014, 2021), we identified the source as being inferential; we also argued that conjectural uses of the FUT have different status in French and in Spanish. In French, they are obtained through pragmatic enrichment processes, linked to future verification of a non-​future situation. As a result, the conjectural FUT reading is strictly limited to specific environments. Conversely, in Spanish, contextual enrichment operates over a less restrictive tense semantics, which does not require future verification. Consequently, conjectural readings of the FUT are much more widespread. In a previous study on the acquisition of FUT (Ahern, Amenós-​ Pons and Guijarro-​Fuentes 2020), no significant differences (p < .05) in the choice of tenses used by L1 Spanish speakers and L1 French learners (at B2 and C1 levels of proficiency in Spanish) regarding the expression of futurity were found. The study showed that the French participants did clearly distinguish between chronological and conjectural interpretations of Spanish FUT (including concessive environments), even though they avoided using the tense to express conjecture. The research reported herein refines the previous results, showing that the L1 French speakers, at the proficiency levels of Spanish being considered, do not acquire the aspectual restrictions of the conjectural FUT. This difference between comprehension and acquisition is fully compatible with the semantic hypotheses that we adopted in Section 2. By paying attention to contextual cues, and transferring L1 knowledge, the learners can infer that, in some environments, FUT does not relate to future time reference. However, our participants do not

196

Aoife Ahern et al.

interiorize the aspectual configurations that limit the conjectural interpretations of the tense. In the case of L1 French speakers, learning the use of the Spanish FUT to express chronology (and understanding its alternation with other means of expressing futurity) is clearly favored by the cross-​linguistically shared features of the tense; conversely, the ability to extend the use of the Spanish FUT to conjectural and concessive environments is hindered by language-​specific feature differences. Therefore, our data suggest that the expression of conjecture by means of the FUT in French comes as a result of contextual enrichment processes, strictly limited by the [+​Forward] feature included in the semantics of the tense. Contrastingly, the [-​Forward] feature of the Spanish FUT makes it more compatible with conjecture and concession. When acquiring the uses of FUT in L2 Spanish, our L1 French learners operate within the limits the interpretative strategies and routines of their own native language. This results in prevalent variability and persistent acquisition difficulties that affect the full command of the linguistic resources for expressing evidentiality, despite the deceptive cross-​linguistic transparency of the Romance FUT.

References Ahern, Aoife /​Amenós-​Pons, José /​Guijarro-​Fuentes, Pedro 2020. Future tense acquisition by French-​speaking learners of L2 Spanish: chronology, conjecture and concession. In Guijarro-​Fuentes Pedro /​Suárez-​Gómez, Cristina (eds) New Trends in Language Acquisition Within the Generative Perspective. Springer: Dordrecht, 27–​48. Azzopardi, Sophie 2017. Le futur est-​il un marqueur modal ? Analyse du fonctionnement du futur à effet de sens conjectural en français et en espagnol. In Baranzini, Laura (ed.), Le futur dans les langues romanes. Berna: Peter Lang, 79–​105. Bertinetto, Pier Marco 2000. The progressive in Romance, as compared with English. In Dahl, Östen (ed.) Tense and Aspect in the Languages of Europe. Berlin: Mouton De Gruyter, 559–​604.

Conjectural Future in L1 French / L2 Spanish

197

Bruhn de Garavito, Joyce 2010. Acquisition of the Spanish plural by French L1 Speakers: The Role of Transfer. In Liceras, Juana /​ Zobl, Helmut /​Goodluck, Helen (eds) The Role of Formal Features in Second Language Acquisition. New York and London: Routledge, 270–​298. Dahl, Östen 2000. The grammar of future time reference in the languages of Europe. In Dhal, Östen (ed) Tense and aspect in the languages of Europe. Berlin: Mouton de Gruyter, 309–​328. De Saussure, Louis /​Morency, Patrick 2012. A cognitive-​pragmatic view of the French epistemic future. French Language Studies, 22, 207–​223. Escandell-​Vidal, María Victoria 2010. Futuro y evidencialidad. Anuario de Lingüística Hispánica, XXVI, 9–​34. Escandell-​Vidal, María Victoria 2014. Evidential futures: The case of Spanish. In De Brabanter, Philippe /​Kissine, Mijail /​Sharifzadeh, Saghie (eds) Future Times, Future Tenses. Oxford: Oxford University Press, 219–​247. Escandell-​Vidal, María Victoria 2021. The simple future in Romance. Semantic parameters and linguistic variation. In Baranzini, Laura /​De Saussure, Louis. (eds) Aspects of Tenses, Modality, and Evidentiality. Cahiers Chronos, 31. Leiden: Brill. Gabriele, Alison 2009. Transfer and transition in the SLA of aspect. Studies in Second Language Acquisition 31, 371–​402. García Negroni, María Marta 2016. Polifonía, evidencialidad citativa y tiempos verbales. Acerca de los usos citativos del futuro morfológico y del futuro perifrástico. In González Ruiz, Ramón /​ Izquierdo Alegría, Dámaso /​Loureda Lamas, Óscar (eds) La evidencialidad en español: teoría y descripción. Madrid: Iberoamericana /​Vervuert, 279–​302. Grüter, Theres /​Crago, Marta 2010. The roles of L1 transfer and processing limitations in the L2 acquisition of French object clitic constructions: evidence from Chinese-​and Spanish-​speaking learners. Proceedings of the Annual Boston University Conference on Language Development 34, 1, 150–​161. Guijarro-​Fuentes, Pedro 2014. On the L2 Spanish acquisition of [+​/​-​] interpretable features. In Costa, Joao et al. (eds), New Directions

198

Aoife Ahern et al.

in the Acquisition of Romance Languages. Cambridge: Cambridge Scholars Publishing, 101–​134. Guijarro-​Fuentes, Pedro 2015. A bidirectional study: Is there any role for transfer in adjective placement? In Guijarro-​Fuentes, Pedro /​ Schmitz, Katrin /​and Müller, Natascha (eds) The Acquisition of French in Multilingual Contexts, Clevedon: Multilingual Matters, 185–​214. Hawkins, Roger /​Hattori, Hajime 2006. Interpretation of English multiple wh-​questions by Japanese speakers, a missing uninterpretable feature account. Second Language Research 22, 269−301. Laca, Brenda /​Falaus, Annamaria 2014. Les formes de l'incertitude. Le futur de conjecture en espagnol et le présomptif futur en roumain. Revue de Linguistique Romane 78, 313–​366. Laca, Brenda 2016. Variación y semántica de los tiempos verbales. El caso del futuro. hal-​01533046f, https://​hal.archi​ves-​ouver​tes.fr/​ hal-​01533​046, accessed 7 January 2022. Lardiere, Donna 2008. Feature-​assembly in second language acquisition. In Liceras, Juana /​Zobl, Helmut /​Goodluck, Helen (eds) Features in Second Language Acquisition. Mahwah, NJ: Lawrence Erlbaum, 106–​140. Lardiere, Donna 2009. Some thoughts on a contrastive analysis of feature in second language acquisition. Second Language Research 25.2, 173–​227. Le Querler, Nicole 1996. Typologie des modalités. Caen: Presses Universitaires de Caen. Matte Bon, Francisco 2007. Las maneras de hablar del futuro en español. Del sistema codificado a las interpretaciones contextuales. MarcoELE 5, , accessed 5 February 2021. Odlin, Terence 1989. Language Transfer. Cambridge: Cambridge University Press. Rocci, Andrea 2000. L’interprétation épistemique du futur en italien et en français: une analyse procédurale. Cahiers de linguistique française 22, 241–​274. Rodríguez Rosique, Susana 2015. Distance, evidentiality and counter-​ argumentation: Concessive future in Spanish. Journal of Pragmatics, 85, 181–​199.

Conjectural Future in L1 French / L2 Spanish

199

Rojo, Guillermo /​Veiga, Aleixandre 1999. El tiempo verbal. Los tiempos simples. In: Bosque, Ignacio /​Demonte,Violeta (eds) Gramática descriptiva del español, volume 2. Madrid: Espasa, 2867–​2934. Schwartz, Bonnie /​Sprouse, Rex 2000. When syntactic theories evolve. In Archibald, John (ed) Second language acquisition and linguistic theory. Oxford: Blackwell Publishers, 156–​186. Sharwood-​ Smith, Michael /​Kellerman, Eric 1986. Crosslinguistic influence in second language acquisition: An introduction. In Sharwood-​Smith, Michael /​Kellerman, Eric (eds) Crosslinguistic influence in second language acquisition. Oxford, UK: Pergamon Press, 1–​9. Sorace, Antonella 2011. Pinning down the concept of ‘interface’ in bilingualism. Linguistic Approaches to Bilingualism 1/​1, 1–​33. Sorace, Antonella 2012. Pinning down the concept of interface in bilingualism: A reply to peer commentaries. Linguistic Approaches to Bilingualism 2/​2, 209–​216. Soto, Guillermo 2008. Sobre el llamado futuro de probabilidad. Boletín de Filología de la Universidad de Chile 43, 193–​206. Squartini, Mario 2001. The internal structure of evidentiality in Romance. Studies in Language 25/​2, 297–​334. Squartini, Mario 2004. Disentangling evidentiality and epistemic modality in Romance. Lingua 114, 83–​895. Squartini, Mario 2008. Lexical vs. grammatical evidentiality in French and Italian. Linguistics 46/​5, 917–​947. Tasmowski, Liliana /​Dendale, Patrick 1998. Must/​will and doit/​futur simple as epistemic modal markers. Semantic value and restrictions of use. In Van der Auwera, Johan (ed.), English as a Human Language. To honour Louis Goosens. Munich, Lincom Europa, 325–​336. Tsimpli, Ianthi M., and Dimitrakopoulou, Maria 2007. The Interpretability Hypothesis: Evidence from Wh-​Interrogatives in Second Language Acquisition. Second Language Research 23, 215−242. Weinreich, Uriel 1953. Languages in contact. The Hague, NL: Mouton. White, Lydia 2003. Second Language Acquisition and Universal Grammar. New York: Cambridge University Press. Wilmet, Marc 1996. Grammaire critique du français. Louvain-​la-​ Neuve : Duculot.

200

Aoife Ahern et al.

Appendix I. Sample items according to the different conditions tested (see t­ ables 2 and 3)

Task 1 item examples Item conditions Sample item Conjectural with (Conversación telefónica entre dos periodistas en un programa stative (1, 2) de radio) Tengo delante de mí un ejemplar del diario La Vanguardia con un titular enorme, que supongo que SERÁ el mismo que tienes tú en la mano en este momento. (Telephone conversation between two journalists on a radio programme) I have in front of me a copy of the La Vanguardia newspaper with an enormous headline, which I suppose WILL BE the same as the one you have in your hands just now. Concessive with (Conversación entre dos amigas, sobre el comportamiento de stative (3 to 5) otra amiga común) Ana: Yo creo que María tiene sus razones para actuar así… Rosa: TENDRÁ sus razones, pero se equivoca. (Conversation between Friends about another friend’s behavior) Ana: I think María has her reasons for acting that way. Rosa: She’LL HAVE her reasons, but I think she’s wrong. Conjectural (Conversación entre una pareja en su casa) Mario: He oído reading, activity voces en casa de los López. Qué raro, ¿no? Ana: Creo que (6 to 8) tienen una comida familiar. HABLARÁN con el abuelo, que está muy sordo. (Couple’s conversation at home). Mario: I’ve been hearing voices from the Lopez’ house. Strange, isn’t it? Ana: I think they’re having a family meal together. [They] WILL TALK to the grandfather, who is very deaf. Concessive with (Conversación entre dos amigos sobre la decisión que ha activity (9 to 11) tomado un amigo común) Miguel: Pedro siempre piensa mucho antes de actuar. Por eso siempre toma buenas decisiones. Antonio: PENSARÁ mucho, pero me parece que esta vez se equivoca. (Conversation between friends about another friend’s decision) Miguel: Pedro always thinks al lot before acting. That’s why he always makes good decisions. Antonio: He’LL THINK a lot, but this time it seems he’s making a mistake.

Conjectural Future in L1 French / L2 Spanish Conjectural with telic (12 to 15)

201

(Conversación entre participantes en un concurso de canto) Rosa: Le han dado el premio a Luis porque lo ha hecho muy bien. José: Le DARÁN el premio a Luis, pero yo creo que María lo ha hecho mejor. (Conversation among the participants in a song contest) Rosa: They gave Luis the prize because he did really well. José: They’LL GIVE Luis the prize, but I think María was better. Concessive with (Conversación entre los profesores de una escuela que han telic (16 to 19) organizado una excursión) Manuel: Luis viene a la excursión. Su madre ha pagado. Ana: La madre PAGARÁ la inscripción, pero ya no quedan plazas. Por eso, no podemos aceptarle. (Conversation between teachers who have organized an outing). Manuel: Luis is coming on the outing. His mum’s paid. Ana: His mum’LL PAY the fee, but there are no places left. That’s why we can’t take him. Task 2 item conditions Item condition Sample item Simple FUT, Supongo que el coche ESTARÁ en malas condiciones después stative, de tanto tiempo… conjecture I suppose the car WILL BE in bad condition after all that time… Simple FUT, Yo creo que ROBARÁN la casa… porque veo a un ladrón que activity, se escapa por la ventana. conjecture (items I think they’LL ROB the house…because I can see a burglar 26, 34, 42) escaping out the window. Simple FUT, Bueno, el coche lo ABANDONARÁN en este parking hace telic, conjecture nueve años, por falta de espacio en el garaje de casa… (items 27, 35, 43) Well, they WILL ABANDON the car in this carpark nine years ago, as they lacked space in their home carpark Compound FUT, Me imagino que HABRÁN ORGANIZADO una fiesta de telic (items 28, cumpleaños… 36, 45) I suppose they’LL HAVE ORGANIZED a birthday party… Compound HABRÁN BUSCADO algo de valor que no encontraban, y por FUT, activity, eso han abierto todos los cajones. conjecture (items They’LL HAVE SEARCHED for valuables they didn’t find, so 29, 37, 46) they opened all the drawers. Compound FUT, Me imagino que el dueño HABRÁ TENIDO algún state, conjecture problema grave… (items 30, 37, 48) I imagine the owner WILL HAVE HAD some severe problem.

Cecilia Mihaela Popescu and Oana Adriana Du ȚĂ

Romance Future: between Epistemic Modality and Evidentiality

Abstract Our paper focuses on a special type of use of “epistemic future” (hereinafter: FUT) in Italian, Spanish and Romanian, in adversative or concessive syntactic and discursive matrices. Relevant scientific contributions (Squartini 2001, 2004, 2005, 2012; Rodríguez Rosique 2015; Zafiu 2002, 2009, 2018, inter alia) propose various interpretations of FUT forms in such contexts, ranging from a purely epistemic reading to an evidential-​ reportative one. Considering this situation, our approach aims at proving that a complex rhetorical strategy is involved in such contexts, where the FUT, in fact, only actualizes the prototypical feature [+​placement in a subsequent relation] (“ultériorité”), exclusively marking the relation of subsequence of the hypothesis compared to a previously mentioned state of facts. In the first part of the paper, we aim to perform a comparative analysis of the semantic and functional behavior of FUT in adversative and concessive contexts and other types of epistemic utterances, so as to mainly emphasize that a description of inferential cognitive processes is actualized in both cases. The second part tries to show, by resorting to the Romance typological perspective, that FUT verbal forms involved in concessive and adversative structures do not have an evidential reportative value per se, like the conditional. We are dealing with a par défaut meaning, actualized through the semantic and syntactic framework of the verb morphemes. Keywords: future, presumptive, Romance languages, evidentiality, epistemic modality, concessive adversative contexts

204

Popescu and Duță

1. Introduction Our paper deals with a special type of usage of “epistemic future” (hereinafter: FUT) in Italian, Spanish and Romanian,1 within adversa­ tive or concessive syntactic and discursive matrices (see ­examples 1, 2 and 3 below), fulfilling the semantic and pragmatic function of counter-​ argument for a specifically presented conclusion: (1) (2)

(3)

Avrà un dottorato, ma non è troppo sveglio (Rocci 2000: 248) [He may have a PhD, but he is not too bright]2 Le parecerá una tontería, pero aquello me salvó (Real Academia Española 2010: 448). [It may seem foolish to you, but that’s what saved me] Om semăna/​om fi semănând noi amândouă, dar nici nu se compară (Reinheimer-​R îpeanu 1994: 514). [We may look alike/​we may be looking alike, but it’s not the same thing]

Scientific literature oscillates in the interpretation of FUT forms in such contexts, between a purely epistemic reading, an evidential-​reportative reading and an illocutionary, attenuative reading (for details on the three types of meanings –​epistemic, evidential-​ reportative3 and illocutionary –​see Squartini 2001, 2004, 2005, 2012; Rodríguez Rosique 2015; Zafiu 2002, 2009, 2018, inter alia). Actually, in all these cases, FUT/​presumptive/​Spanish conditional

1 2 3

The Romanian epistemic FUT is referred to as “presumptive” by the grammars of this linguistic system. Translation by the authors of the paper, unless otherwise indicated. By ‘evidentiality’ we understand, as proposed by Chafe and Nichols (1986) or by de Haan (1999, apud Brugman and Macaulay, 2015), a semantic archi-​category that implies the discursive marking by the speaker of the source lying at the basis of the information conveyed in the utterance. The meaning of ‘epistemic modality’ translates the degree of certainty awarded by the speaker to the propositional content of the utterance and, implicitly, a certain propositional attitude of the speaker regarding his/​her own utterance, which s/​he qualifies as or (see Palmer 1986: 10–​11).

Romance Future - Epistemic Modality and Indirect Evidentiality

205

(hereinafter: COND) forms are generally classified as operators of epistemic (subjective) modality expressing a “conjecture” argued by the speaker regarding a certain situation mentioned “somewhere” in the sentence. A recent different interpretation considers that, in such sentences, the concerned verbal paradigms are evidential markers showing that the source of the asserted information is indirect, reportative. Finally, a third point of view argues that FUT/​presumptive/​ COND is used in these concessive/​adversative structures with an illocutionary, attenuative value. As for this last type of interpretation, it should be said (the discussion will not be further developed) that FUT forms are not (generally) used very frequently with such an attenuative value in Romance languages (especially in Romanian –​where the canonical FUT, i.e. voi cânta [I shall sing], is preferred instead of presumptive in attenuative uses). Against this background, our approach aims to show that a complex rhetorical strategy is involved in such contexts, with the global significance of describing cognitive, inferential processes (the speaker provides a strongly subjective, almost speculative, assessment of a (potential) situation) and, at the same time, of expressing the primary source of information –​indirect, reportative evidentiality. However, we consider that, in such contexts, FUT actually only expresses the prototypical feature [+​placement in posteriority (‘ultériorité’)],4 exclusively marking the posteriority of the hypothesis compared to a previously stated situation (see Popescu 2013, 2015b). 4

By ‘posteriority’ we refer to the concept of ‘ultériorité’, coined by Gosselin (2001) and also deployed by Vetters (2001) and Brès (2012) whose value is not limited to strictly expressing a temporal, chronological, syntactic relation. In our case, “ce terme a l’avantage par rapport à futur qu’il n’exprime pas seulement la postériorité par rapport au présent de m0 [monde actuel], mais par rapport à n’importe quel moment dans n’importe quel monde –​réel ou imaginaire –​[…]” [this term presents the advantage, compared to future, that it does not solely express posteriority in relation to the present of m0 [today’s world], but compared to whatever moment in whatever world –​whether real or imaginary –​ (…)] (Vetters 2001: 177). Furthermore, “l’ultériorité pourrait s’expliquer aussi par le fait que la signification de la prédication au FUT (pFUT) est calculée à partir de la relation d’implication [plutôt argumentative et pragmatique que logico-​temporelle –​ cf. ‘a temporal pragmatic implication’, chez Gennari (2000, 268)] qui s’établit entre l’ensemble référentiel R et le moment T0 de la situation

206

Popescu and Duță

In the first part of our paper, we aim to provide a parallel analysis of the semantic and functional behavior of FUT in adversative and concessive contexts and in other types of epistemic utterances, so as to emphasize that a description of cognitive, inferential processes is provided in both cases. However, we shall see that the situation described in conversational matrices with a purely epistemic value has a strongly uncertain nature. From an evidential perspective, such structures express indirect, inferential evidentiality and the speaker is an active participant to the asserted content. Adversative or concessive structures are, instead, neutral from a truth-​conditional point of view, as the analyzed verbal morpheme only provides an explicit description of the modality of neutral possibility (i.e. the degree of commitment to the truth of the proposition is not specific), like the equivalent French structures il se peut que/​peut-​être que [it is possible that/​possibly/​perhaps]. Therefore, such adversative and concessive matrices involve indirect, reportative evidentiality, with the speaker taking a certain distance from the content of the asserted utterance. The second part of the paper attempts to show, by resorting to the Romance typological perspective, that FUT verbal forms involved in concessive/​adversative contexts do not have an evidential-​reportative value by themselves, as COND does. This is a by default significance activated by the semantic and syntactic framework of the analyzed verbal morphemes. They only inherently express the cognitive inferential d’énonciation (=​Sit0). C’est grâce à cette perspective étendue sur l’ensemble référentiel R et à la relation spécifique de R avec le moment T0 que ce décalage temporel d’ultériorité peut être dénoté. Il est opéré par transfert cognitif dans le cas des emplois modaux épistémiques, ou bien, par une réévaluation illocutoire, dans le cas des emplois illocutionnaires, atténuatifs” (Popescu 2015b: 61– 62) [posteriority could also be explained by the fact that the meaning of the predication in FUT (pFUT) is determined based on the relation of implication [rather argumentative and pragmatic than logical and temporal –​‘a temporal pragmatic implication’, according to Gennari (2000, 268)] established between the R ref­ erential complex and the T0 moment of the utterance (=​Sit0). This extended perspective on the R referential complex and to the specific relation between R and the moment T0 makes it possible to express this temporal delay of posteriority. This is done by means of cognitive transfer in the case of epistemic modal uses or by means of an illocutionary revaluation in the case of illocutionary, attenuative uses].

Romance Future - Epistemic Modality and Indirect Evidentiality

207

context having a relation of posteriority compared to a referential structure (R), subjectively assessed by the speaker and holding a relation of posteriority of the hypothesis through predication in FUT. This R may represent a true situation –​as in (1), “Lui ha un dottorato” [He has a PhD], a “discursive truth” –​as in (3), where the predication “Om semăna noi” [We may look alike] is a meta-​discursive representation of the sequence: “You say/​it is said: we look alike”, in fact, the representation of a secondary speech act.5

2. Epistemic Future vs Concessive Future 2.1. General Features of Epistemic Future in Romance Languages In Italian, Spanish, French and Romanian, the epistemic FUT (also referred to as conjectural, probability, suppositional, inferential or putative FUT in scientific literature) is seen in specific conversational matrices, generally assertive-​affirmative (­examples 4, 5a, 6 and 7 below) or dubitative patterns (5b): (4)

Pourquoi donc a-​t-​on sonné la cloche des morts? Ah! mon Dieu, ce sera pour Madame Rousseau (the classical example quoted by Imbs (1968) and reproduced in Rocci (2000: 244)). [So why did the dead bell ring? Ah! My God, it could be for Madame Rousseau]

5 By secondary speech act, we understand another speech act q, on which is based the speech act p. For exemple, in (3), the speech act p “Om semăna noi” [We may look alike] is based on the sequence: “You say/​it is said: we look alike”, which is a secondary speech act q. More precisely, a secondary speech act represents the sentence produced by an original speaker, which could be a human being, but also the public rumor.

208

Popescu and Duță

(5)

a. Acabo de ver las luces. Juan ya estará en casa (Gennari 2000: 264). [I’ve just seen the lights. Juan must be at home already] b. ¿Estará enfadado conmigo? (Squartini 2004: 73). [Could he be angry with me?]

(6)

A: non c’è Marco? E: e_​sarà sarà in ospedale e’diventato papa’oggi (Corpus de conversations du LIP, MA23: 75–​80, apud Rocci 2000: 245). [A: Marco’s not here? E: He could, could be at the hospital, he’s become a father today]

(7)

Petru nu e acasă. O veni mai târziu. [Petru is not home. He may come later]

As shown on various occasions (see Popescu 2013: 50, Popescu 2015b: 60), the properties of such occurrences may be described in terms of three factors: –

– –

the strongly subjective attitude of the speaker regarding the asserted content. S/​he expresses his/​her lack of trust (in an epistemic meaning) regarding the truth of his/​her utterance, showing that s/​he does not have all the information/​k nowledge to allow him/​her to assert it with full certainty; the described situation is not future, but present or, at most, a future close to present (see e­ xample 7 above); the degree of certainty awarded by the speaker to the predication is of strong probability, with the meaning of such structures being close to utterances explicitly modified by modal operators such as (Rom.) probabil că, (It.) probabilmente/​forse che, (Fr.) probablement, etc., used with the present indicative. This degree of knowledge expressed by the speaker is obtained through inference.

Our opinion is that, in such epistemic contexts, FUT only expresses the prototypical feature [+​prospective placement] or, more specifically, [+​placement in posteriority (“ultériorité”)], since this verbal morpheme only marks a posteriority of the hypothesis compared to the previously assessed situation. It is the content of the hypothesis that

Romance Future - Epistemic Modality and Indirect Evidentiality

209

is chronologically represented in T0, not this relation of posteriority. In other words, the verbal form only expresses the relation of posteriority/​ulteriority that exclusively describes the cognitive process, while the content of the hypothesis is placed in the present(-​future) frame of the speaker, as this hypothesis is linguistically performed through a (declarative or interrogative) speech act and “les actes de langage sont par définition exécutés par le locuteur au moment de l’énonciation” (Vet/​Kampers-​Manhe 2001: 96) [speech acts are, by definition, per­ formed by the speaker as s/​he utters them]. Before ending the discussion on the particularities of the occurrence of Romance epistemic FUT, two further remarks should be provided on other parameters frequently considered by scientific literature when defining this type of operation of FUT. For instance, it is frequently said that the use of this verbal morpheme in purely epistemic contexts is due to the fact that the event/​process affected by the verbal form of FUT will be validated in a prospective frame. In our opinion, neither the analyzed morpheme, nor the overall meaning of the utterance suggest (explicitly or implicitly) the idea of future validation of the event p. Moreover, if it is considered that the FUT verbal form only describes a cognitive procedure in such contexts, without insisting on its consequences, then, a future validation of the truthfulness of the concerned utterance does not seem necessary. Furthermore, many authors claim that the predication marked by a Romance form of epistemic FUT should be assigned a value of uncertain truth. Regarding this, we consider that such an attribute applies rather to the denoted situation, not to the (cognitive) process described by the FUT verbal morpheme. 2.2. General Features of Concessive Future/​Presumptive in Romance Languages On the other hand, the epistemic value of FUT is not only seen in assertive-​affirmative or dubitative patterns but also in a special adversative or concessive conversational matrix, where the FUT verbal paradigm also has an evidential-​reportative value according to scientific literature (see Zafiu 2009: 289–​305). It is the so-​called concessive

210

Popescu and Duță

FUT which is found, except in French,6 in syntactic structures such as those in e­ xamples 1, (2 reproduced in 9) and 3, as well as in 8 and 10: (8)

Sarà, ma non ci credo (Rocci 2000: 248). [It may be like this, but I don’t believe it]

(9)

Le parecerá una tontería, pero aquello me salvó (Real Academia Española 2010: 448). [It may seem foolish to you, but that’s what saved me]

(10) O fi el frumos, dar mie tot nu-​mi place. (Popescu 2018: 381) [He may be handsome, but I still don’t like him.]

Such adversative and concessive structures, whose discursive function is that of counter-​argument for a conclusion that is now explicitly presented, denote several possibilities: –

– –

6

7

a subjective attitude of the speaker, who “[...] peut varier assez librement entre les deux extrêmes de la réfutation et de la concession pleine” (Rocci 2000: 248) [can oscillate quite freely between the two extremes of refusal and full agreement]. a situation that is not future, but present, or even present-​future (see e­ xample 9).7 a degree of knowledge of the asserted propositional content that may range from strong probability to possibility or even full certainty, as there are indeed utterances whose truth value “se trouve être présupposée [pour vraie] dans le contexte d’énonciation” (Rocci 2000: 248) [is presupposed [as true] in the context of the utterance], as in (1 reproduced in 11), where the validity of the predication “Lui ha un dottorato” [He has a PhD] is well known by the participants in the exchange and may be validated as [certain/​true].

It should be said that concessive FUT is seldom found in French, and only in popular language; explicitly modalised structures are used otherwise: il se peut que /​peut-​être que. An analysis would be needed to see whether temporality does not result, in “modal” uses, from reasons related to the internal aspect of a verb (Aktionsart).

Romance Future - Epistemic Modality and Indirect Evidentiality

(11)

211

Avrà un dottorato, ma non è troppo sveglio (Rocci 2000: 248) [He may have a PhD, but he is not too bright]

Against this background, we consider that the meaning of concessive FUT should not be understood as changing the truth value of the utterance, but rather still as a marker of cognitive inferential processes, suggesting a strongly subjective, almost speculative assessment of a situation (ranging from close to certainty to potentiality). For this reason, all these examples of concessive FUT in Italian, Spanish or Romanian are translated into French using the phrase il se peut que /​ peut-​être que […], which performs the same function as the analyzed morpheme, i.e. providing an explicit description of the modality of neutral possibility. This interpretation of concessive FUT, as a discursive mechanism to describe certain cognitive processes, may also explain another value that has been recently assigned to it in scientific literature, i.e. the “evidential” or, more precisely, “reportative” one (see also Zafiu 2018: 116–​127).8 In other words, in any concessive/​adversative sentence built with a form of FUT (or presumptive), the source of information refers to the words of a different person (frequently identified, but not always, as the receiver). Thus, a sentence such as (10) may be paraphrased as below in (12): (12)

8

[Se (lumea) zice/​t u zici] că e frumos, dar mie tot nu-​mi place. [[People say /​you say] he is beautiful, but I still don’t like him]

It should be said that, in Portuguese, the FUT form is used with a clear, explicit evidential-​reportative value. In such cases, its French, Italian or Romanian equivalent includes (present or past) COND (see Squartini 2005: 249): A rádio estatal da Serra Leoa confirmou a morte de um dirigente rebelde, o capitão Solomon Musa, que foi primeiro-​ministro da junta militar […]. Musa terá sido morto a 29 de Dezembro, em Benguema (Diário de Notícias 1.2.1999, p. 14, apud Squartini 2005: 249) [The Sierra Leone national radio confirmed the death of a rebel leader, captain Solomon Musa, who was the prime minister of the military junta […]. Musa probably died on December 29 in Benguema].

212

Popescu and Duță

However, this content is always subject to the speaker’s (even speculative) assessment, which can again suggest an incompatibility between concessive FUT and the structures where the source of information explicitly refers to a completely different person (see 13a, b):9 (13)

a. *Potrivit ANM, la munte o ploua. [According to the weather forecast, it may rain in the mountains] b. *Ion spune că o ploua. [John says it may rain]

3. A Typology of Discursive Manifestation of Concessive Future on an Inter-​Romance Level After the attempt of providing a noematic definition of concessive FUT in the paradigm of epistemic modality and, especially, in that of evidentiality, we now aim to provide a brief description of how this concessive/​ adversative pattern is expressed in Spanish, Italian and Romanian. The purpose is to outline a typology of manifestation on an inter-​Romance level, as well as to provide new data supporting the possibility to decrypt concessive/​adversative structures in the light of recent evidentiality theories. 3.1. The Manifestation of Concessive Future in Spanish In modern continental Spanish, COND (the –​ría form) is used as a “duplicate” of FUT forms both in purely epistemic structures and in concessive/​

9

In such structures where the source of information explicitly refers to a com­ pletely different person, presumptive forms are not accepted (as seen in 13a and 13b). Only COND forms are accepted because they mark citational evidentiality and a certain distance of the speaker from the asserted content.

Romance Future - Epistemic Modality and Indirect Evidentiality

213

adversative patterns, as seen in the comparison of the examples below (cf. the FUT form in 14a with the COND forms in 14b, c and d): (14)

a. Le parecerá una tontería, pero aquello me salvó (Real Academia Española 2010: 448) [It may seem foolish to you, but that’s what saved me] –​concessive FUT, referring to the present. b. Sería fea, pero era simpática [She might be ugly, but she was cute] –​the simple –​ría form with a concessive value and reference to the past. It can be translated into Romanian by a past presumptive form: ‘o fi fost urâtă/​poate că era urâtă, dar era simpatică’ [she might be ugly/​maybe she was ugly, but she was cute] (Reinheimer-​R îpeanu 2001: 289). c. Lo habría hecho, pero yo no lo creo [He might have done it, but I don’t think so] –​the compound –​ría form with a concessive value and reference to the past can be translated into Romanian by a past presumptive form as well: ‘o fi făcut-​o, dar eu nu cred’ (Reinheimer-​ Rîpeanu 2001: 289). d. Muy bondadoso sería, pero bien podía ser absolutamente imaginario (Real Academia Española 2010: 450). [He might be very kind-​hearted, but it could simply be a figment of my imagination] –​the simple –​ría form with a modal, concessive value, related to a moment of reference in the past can be translated into Romanian by a past presumptive form: ‘Poate că o fi fost el foarte bun la suflet, dar putea fi foarte bine și doar rodul imaginației lui’.

In all these examples, the two verbal Spanish forms, FUT and concessive COND, express a highly subjective assessment, close to speculation, of a (potential) situation and, as discursive mechanisms describing certain cognitive processes, they may have an “evidential” or “reportative” value (see 14c). Moreover, it has to be emphasized that the replacement of synthetic FUT by the periphrasis ir a +​infinitive also changes the meaning of the utterance, from a strongly probable potential or a possible potential connected to the present to marking weak conjecture in the future in such discursive structures (see 15a and 15b below): (15)

a. Va a cerrar la puerta, pero no lo conseguirá. [He is going to close the door, but he won’t make it]

214

Popescu and Duță b. Cerrará la puerta, pero no lo conseguirá (Jensen 2002: 122) [He will close the door, but he won’t make it]

Furthermore, according to Jensen (2002: 122): […] la diferencia entre los dos está en que ir +​infinitivo indica que, a juzgar por lo que se ve y sabe, se realizará la situación (en sentido prospectivo), mientras que el futuro se interpreta como solamente indicando la opinión subjetiva del hablante acerca de una situación futura. (Jensen 2002: 122) [[…] the difference between [these two forms] lies in the fact that ir +​infinitivo shows that, considering things that are seen and known, the situation will take place (in a prospective sense), while the future is interpreted as only indicating the speaker’s subjective opinion on a future situation]

3.2. The Manifestation of Concessive Future in Italian Contemporary Italian is the linguistic system where adversative/​concessive discursive structures are perhaps most widely used, and they are here exclusively constructed with FUT forms (not COND). We are dealing with syntactically independent sentences, compulsorily coordinated to another sentence with ma or però (see 8, reproduced below in 16): (16)

Sarà, ma non ci credo (Rocci 2000: 248). [It may be like this, but I don’t believe it]

As in Spanish or Romanian (see 3.3), the FUT morpheme expresses a wide range of degrees of epistemic truthfulness in such cases: generally, a strongly probable potential (as in (17)), which may even come close to certainty (see 18), as well as possible potential –​ see (19) and (20) –​and has, as seen in Spanish, a strong (explicit) anchorage in the area of evidentiality: (17)

Tu riuscirai anche a battterlo, non lo nego; ma lui gioca decisamente meglio (Renzi et al. 1991, II: 115–​121).

Romance Future - Epistemic Modality and Indirect Evidentiality

215

[You will even manage to defeat him, I don’t deny that; but he definitely plays better] (18)

Avrà un dottorato, ma non è troppo sveglio (Rocci 2000: 248) [He may have a PhD, but he is not too bright]

(19)

Non dirò che fossi soddisfatto, però non protestai (Renzi et al. 1991, II: 115–​ 121) [I won’t say I was satisfied, but I did not protest]

(20)

La cosa più bella sembra lo schermo: finalmente via lo sfondo così nero che sarà anche chic ma alla fine stufa (“La Repubblica” 16 giugno 2010) http://​w ww.treccani.it/​enciclopedia/​f uturo_​(Enciclopedia-​dell’Italiano) [The screen seems to be the best thing: they have finally got rid of the black background, which may be chic, but, in the end, is so boring]

The interesting fact in Italian is that FUT is placed in opposition to COND forms (unlike in Spanish), with the first verbal paradigm marking the hypothesis expressed by the speaker (see 21), and the second one referring to reportative evidentiality (see 22) (see Squartini 2012: 2118). This distinction reinforces the idea that, within this linguistic system (and not only, also see 3.3), FUT is still a strong marker of evidentiality obtained through deductive, inductive or abductive inference. (21)

Secondo me, Andrea sarà già arrivato (see Squartini 2012: 2118) [According to me, Andrea will have already arrived]

(22) Secondo Marco, Andrea sarebbe già arrivato (see Squartini 2012: 2118) [According to Marco, Andrea might have already arrived]

3.3. The Manifestation of Concessive Future in Romanian Finally, in Romanian, the adversative/​concessive structures that suggest weakly probable potential, even close to the upper limit of possible potential (see Popescu 2015a), include, as already seen, a verbal form of presumptive with a FUT basis [first, the homonymous form with apheresized FUT, oi cânta (=​Foi) ‘I may sing’; second, oi +​fi +​gerund of the lexical verb, oi fi cântând (=​FoiG) ‘I may be singing’; third, oi +​

216

Popescu and Duță

fi +​past participle of the lexical verb, oi fi cântat (=​FoiP, in this latter case, for reference to the past) ‘I may have sung’].10 Foi and FoiG may be translated into Italian and Spanish by concessive FUT (the modalising adverb quizá ‘maybe’ with subjunctive may also be used in Spanish –​as in (23)), and into French by means of the explicitly modalised phrases il se peut que/​peut-​être que […] (see 23 and its translations and 24) or by means of a concessive subordinate clause. (23)

Om semăna/​om fi semănând noi amândouă, dar nici nu se compară (Reinheimer-​R îpeanu 1994: 514) –​weakly probable potential expressed by Foi/​FoiG referred to the present, in adversative/​concessive structures. fr. “Il se peut que nous soyons semblables, mais la comparaison est hors de doute” it. “Ci si somiglierà anche, ma non si può paragonare” (Squartini 2005: 252). sp. “Quizá nos parezcamos nosotras, pero no hay lugar para la comparación”. [We may look alike/​we may be looking alike, but it’s not the same thing]

(24)

a. Om învăţa noi, dar tot nu vom reuşi să luăm marele premiu (Tuţescu 2007: 561) [We may study, but we still won’t manage to get the great prize] b. O fi el departe şi îngropat în bani, dar informaţiile sunt corecte (EVZ, 3345/​2003, 1 apud Zafiu 2009: 301) [He may be far and swimming in his own money, but the information is accurate]

10 In Romanian, the presumptive mood has three forms (two for reference to the present and one for reference to the past): (i) a special structural type of its four FUT forms, oi cânta (i.e. the apheresised form of the auxiliary a vrea < Lat. velle, in the future tense and the short infinitive of the lexical verb) ‘I may sing’ (acronym: Foi) with reference to the present, and (ii) a periphrastic structure including the future tense of the auxiliary a fi ‘to be FUT’ and the gerund of the lexical verb for expressing present tense, or past participle for expressing past tense: -​(v) oi fi cântând ‘I may be singing’ (acronym: FoiG) –​for reference to the present; -​(v) oi fi cântat ‘I may have sung’ (acronym: FoiP) –​for reference to the past (see also Popescu/​Duță 2017: 33).

Romance Future - Epistemic Modality and Indirect Evidentiality

217

In fact, it should be stated from the very beginning that, in these concessive structures, the analyzed verbal forms may be replaced –​ though more seldom –​by the other three forms of FUT tense in Romanian, as seen in (25a, b and c):11 (25) a. Vom învăţa noi, dar tot nu vom reuşi să luăm marele premiu. b. O să învăţăm noi, dar tot nu vom reuşi/​o să reuşim să luăm marele premiu. c. Avem să învăţăm noi, dar tot nu vom reuşi să luăm marele premiu. [We shall learn, but we still won’t manage to get the big prize]

The examples in (25) show that the deictic FUT (canonical FUT, vom învăța ‘we shall learn’ (25a)), or popular periphrastic structures 11 We remind that, in contemporary Romanian, FUT has four verbal paradigms (which is a particularity of Romanian language in the Romance array), whose morphematic structure is presented below: FUT –​TYPE 1 is considered to be the exclusively literary canonical form of future, “with a certain scholar shade” (see Avram 1997 apud Reinheimer-​ Rîpeanu 2001: 270): a vrea ( generic inferential evidentiality > conjectural inferential evidentiality

Additionally, when the conjectural meaning arises, the inferential component becomes less significant and the subjective presence of the speaker more important in the constructional meaning of “deure +​ infinitive”. In fact, the conjectural meaning is crucially found in interrogative sentences, as in (15–​18). (15)

Ell. ('hont lo deu portá 'l diná?) (Verdú, Josep: ¡Tanner!, CTILC, 1880) [He. (where must he bring the lunch?)]

(16)

I per què, Déu meu! Per què la devien tenir tancada, i tant tancada!, aquella capsa de musica? (Rusiñol, Santiago (Jordi de Peracamps): El poble gris, p. 290, CTILC, 1902) [And why, Oh My God, why would they have it closed, and so closed, that music box?]

(17)

es Rectó m'envia qu'ey vayes que t'ha mesté. Roagó. ¿Qu'em deu volé fer obré? (Mut, Tomàs: Entremes d'en Roagó Florit y na Faldó, CTILC, 1846) [the priest sent me to tell you that he needs you. Roagó: ‘Will he want to make me a worker?]

(18)

¿Qué hi déu havé á las rajolas que las mira tant atent? (Roure, Conrad: Lo castell y la masia, CTILC, 1891) [What will there be in the floor that he is looking at it with so much attention?]

Unlike specific or generic inferences found in Old and Modern Catalan, the conjectural meaning rising in Contemporary Catalan

The Modal deure (‘must’) in Contemporary Catalan

263

is subjectively construed. That would explain why it can be found in interrogative sentences (as well as in declarative and exclamative ones). As a matter of fact, 25/​135 tokens with an evidential reading are part of interrogative sentences and have a conjectural meaning. A small reconfiguration of modality and evidentiality may happen in modal auxiliaries in Contemporary Catalan. It seems that the core meaning of deure is generic and conjectural inference and loses specific inference (closer to root modality), which is expressed by haver de (Sentí 2017). In fact, in spoken present-​day Catalan the situation remains more or less the same, according to our data from the Corpus Oral Dialectal (Spoken Dialectal Corpus, COD) (Sentí 2022). This corpus is not really strong (it is made of 32 monologues); in fact, we just found 21 tokens of modal verb deure. Only 2/​21 tokens of deure convey a deontic meaning, both in Valencian speakers. The other 19/​21 tokens convey an inferential meaning. In most of tokens, generic inference and conjecture prevail. Another argument can be still given in favor of this inferential gradient. Cross-​linguistically, deure counterparts in Romance and Germanic languages display a different set of pragmatic values. French devoir, Italian dovere, German müssen and Dutch moeten express specific and generic inference, but not conjecture. On the other hand, English must covers the three inferential values (see Squartini 2008 for Romance and Mortelmans 2012 for Germanic). According to our data, it seems that Catalan deure resembles English must. In fact, it is closer to American English, since must has lost its root use (deontic meaning) and it is expressed by have to (Abraham 2002: 20–​22),9 as in most of Catalan. It could be linked, though, the loss of the root use (deontic) and the extension to conjectural meanings (inferential).

9

We would like to thank the anonymous reviewer 2 for this observation about American English.

264

Andreu Sentí

Table 6.  Inferential values in Romance and Germanic languages.

devoir dovere deure müssen moeten must

Specific inference

Generic inference

✓ ✓ ✓ ✓ ✓ ✓

✓ ✓ ✓ ✓ ✓ ✓

Conjecture

✓

✓

The inferential gradient seems to be significant in diachrony and language comparison. The new value, conjecture, is more subjective than the previous inferential nuances (specific and generic) and the inferential component becomes less relevant (at least less prototypically inferential) because there is no more external evidence. In fact, it is more dubitative (Squartini 2008). That may indicate that conjecture is closer to epistemicity than evidentiality.

4. Conjecture: An Evidential or an Epistemic Value? We consider modality and evidentiality different linguistic categories (Aikhenvald 2004). However, inferential evidentiality and epistemic modality are two notions which in many languages are very close. This has given rise to a debate on the relation between these categories (see de Haan 1999, Plungian 2001, Squartini 2008, Cornillie 2009, Cornil­­­ lie, Marín-​Arrese & Wiemer 2015, a. o.). Traditionally, the non-​deontic reading of modal verbs such as deure has been qualified as epistemic (Gavarró & Laca 2002).10 However, there are several studies supporting that the essential element in these constructions is actually of an evidential type. According to Squartini (2008: 922), even in those approaches 10 The cognates of deure in other languages are analyzed as epistemic or evidential by different scholars (Dendale 1994, Diewald 1999, 2000, Goossens 2003, Mor­­­­­ telmans 2000, Cornillie 2007, Squartini 2008).

The Modal deure (‘must’) in Contemporary Catalan

265

that hold that modality and evidentiality overlap (van der Auwera & Plungian 1998) or that the basic value is epistemic (Kronning 1996; see also Barbet 2012), it is considered that the modal verb is related to the source of information. Actually, Squartini (2008) points out that the Italian and French modal verbs dovere and devoir as well as the conjectural future encode inferential values. On the contrary, Dendale (1994) considers devoir to be inferential, but the conjectural future to be epistemic (Dendale 2001). Following Nuyts (2017), epistemic modality and inferential evi­ dentiality are really close meanings because “both concern the reality status of a hypothetical state of affairs” and “both involve a reasoning process”. But these categories should be distinguished: the two categories differ from each other because each one focuses on one particular aspect of the process: i) the inference “refers to the reasoning process as such, and denotes its reliability in view of (the quality of) the source information”; and ii) epistemic modality “codes the result of this reasoning process”. Therefore, if a construction has a given inferential meaning it is not necessary the case that it will also have a predetermined epistemic value. As we will argue below, the modal deure focuses on the reasoning process and, therefore, it is inferential. It is true that conjecture is the closest inferential value to epistemic modality given that there is no external evidence. The speakers indicate a hypothesis not obviously substantiated. Nevertheless, we will consider deure as an evidential device with pragmatic epistemic values. In this regard, following Cornillie (2009), there is no link between a particular evidential value (reportative, sensorial, inferential) and a given derived pragmatic epistemic meaning. The reliability of the source comes from factors not connected with the evidential features (mode of knowing, type of evidence, source of evidence) (Cornillie, Marín-​Arrese & Wiemer 2015, cf. Section 3). The modal deure can be seen as an evidential with a [-​SELF] or [+​SELF] in a continuum of source of evidence. The conjectural meaning is, of course, a case of [+​SELF]. In other words, the evidence comes from the speaker (intuition or thinking). In the examples from our corpus, the conjectural use of deure is frequently found in interrogative contexts, as we have seen above. The speaker uses the construction

266

Andreu Sentí

to convey his/​her hypothesis, as in (19), (20) and (21) –​these examples have been quoted previously and are now repeated here again in order to facilitate our argumentation. In fact, as it has been said above, conjecture is a really subjective value. It seems that it functions as a subjective marker oriented to the speaker: (19)

es Rectó m'envia qu'ey vayes que t'ha mesté. Roagó. ¿Qu'em deu volé fer obré? (Mut, Tomàs: Entremes d'en Roagó Florit y na Faldó, CTILC, 1846) [the priest sent me to tell you that he needs you. Roagó: ‘Will he want to make me a worker?]

(20) Ell. ('hont lo deu portá 'l diná?) (Verdú, Josep: ¡Tanner! CTILC, 1880) [where must he bring the lunch?] (21)

[(Reparant en Roch.)] ¿Qué hi déu havé á las rajolas que las mira tant atent? ¡Misteris y mes (Roure, Conrad (Pau Bunyegas): Lo castell y la masia, CTILC, 1891) [What will there be in the floor that he is looking at it with so much attention?]

One must ask whether these examples are linked to low epistemic commitment due to their subjective construal. Is the speaker showing his/​her epistemic commitment? Actually, in interrogative sentences the use of deure is not necessary because the context is already a context of uncertainty. With polar interrogatives, as in example (19), the speaker makes a conjecture about a whole state of affairs: Will he want to make me a worker? The speaker does not have access to the state of affairs and this is why (s)he asks the addressee the truth of the proposition or to elaborate a hypothesis about this truth. However, there are some differences with wh-​questions.11 For example, in (21), the speaker, instead of asking what is there in the floor that makes someone else look at it with so much attention, he instead uses the construction of deure in an interrogative context. With this use, the speaker asks the addressee to elaborate a hypothesis, he asks what can be inferred, he asks for the inferential process, the reasoning: what can be in the floor that makes

11 We would like to thank the anonymous reviewer 1 for this distinction between polar and wh-​questions.

The Modal deure (‘must’) in Contemporary Catalan

267

someone look at it? The speaker is not asking for the probability or for epistemic commitment. Another example is (16) “Per què la devien tenir tancada, i tant tancada!, aquella capsa de musica?” (‘Why would they have it closed, and so closed, that music box?’) in which the speaker knows the state of affairs (the box was so closed). (S)he is asking the addressee about his/​her view about the reason of having closed that box. Therefore, the modal deure is not downgrading the speaker’s commitment to the truth of the state of affairs, but it is asking about the inferential process that would explain the information gap expressed by the wh-​word. In short, our proposal is that the function of deure is not to codify a value of uncertainty but the speaker’s subjective reasoning and his or her point of view. Actually, it seems that this very same function is performed by the Italian epistemic future. The Italian sentence in (22) would be with deure in Catalan, as can be seen in (23): (22)

Avrà dimenticato l’appuntamento? (Squartini 2018)

(23)

Es deu haver oblidat de la cita? [(lit.) Must (s)he have forgotten the appointment?]

Squartini (2018: § 3) says that “what the speaker is doing by using a Future is not reducing the epistemic commitment, but rather insisting that what is asserted has to be attributed to his/​her internal reasoning”. This is also the case with deure in Catalan, in the sense that the auxiliary is focusing on the reasoning. Both Italian future and Catalan deure are [+​SELF]. However, Squartini (2018: 203) defends an evidential-​epistemic relation: “the speaker, in presenting his/​her own surmises, might also be asserting the truth of the propositional content of the sentence”. Anyway, what deserves to be highlighted here, under our point of view, is the fact that conjecture is not an example of low commitment but on the contrary “what the speaker is doing by using a Future is not reducing the epistemic commitment, but rather insisting that what is asserted has to be attributed to his/​her internal reasoning” (Squartini 2018: § 3).

268

Andreu Sentí

To sum up, the inferential gradient and also the new value arisen in Contemporary Catalan –​conjecture–​can be considered evidential notions. The degree of epistemic commitment would be, then, related to contextual features. There is another factor that reinforces this line of analysis. In our corpus, we often find that deure co-​occurs with several epistemic markers (hedges). As we see in the following examples, we can find different degrees of epistemic commitment. On the one hand, we have per suposat, sens dupte, estich cert, bé or ja which indicate a high degree of commitment, as shown in (24) -​(28), on the other hand, em sembla que, presumo que or ab prou feynes signal a low commitment, as shown in (29)-​(31): (24)

Per suposat que deu volgué un marit, que encara que tingui mal génit (Roure, Conrad (Pau Bunyegas): La casamentera, CTILC, 1894) [Of course she must want a husband, even if he is bad-​tempered]

(25)

pero, sens dupte, degué ocasionarlo algun descuit de la pobre mare. (Anònim: La Xinxa, p. 3, CTILC, 1869) [but, without any doubt, it must have caused it some mistake of my mother]

(26)

Estich cert, com si ho vegés, que tot lo dia deu anar amunt y avall fent camins per acudir á nostre Senyor (CTILC) [I am sure, as if I could see it, that he must go all the day up and down, going through different roads, to go to Our Lord]

(27)

no sé encara las costums de la mateixa; pero de tots modos, bé deu haberhi escepcions. (Roure, Conrad: Passions funestas; 1898; CTILC) [I don’t know their habits, but anyway, there must be truly some exceptions]

(28)

Segons tench entes el general axisteix á dites reunions, ja's dehu devertir molt. (Picó i Campamar, Ramon, CTILC, 1874) [According to what I know, the general will attend the mentioned meetings, he probably must have a lot of fun]

(29)

estigas tranquila, com jo presumo que ho deus estar, segueix cumplint ab els teus debers (Roure, Conrad (Pau Bunyegas): Passions funestes, 1898) [stay calm, as I guess you must be, keep fulfilling your duties]

(30)

dins la rectoría, es molt petita, y ab prou feynes se dèu haver vist may plena, puix de 16,000 habitants no hi há més que 60 catòlichs (Verdaguer, Jacint: Dietari d'un pelegrí á Terra Santa, p. 112, CTILC, 1889)

The Modal deure (‘must’) in Contemporary Catalan

269

[inside the parish house, which is really small, and it must have hardly ever looked full, since out of 16.000 inhabitants there are no more than 60 Catholics] (31)

tocaré 'l dos, perque 'm sembla que la Pauleta ja deu estar ab ánsia. [(Va per sortir del llit.)] Can· ¿Que fa?... (Pous, Josep Maria: ¡Mala nit!, CTILC, 1888) [I’ll leave, because it seems to me that Paulete must already be anxious]

The heterogeneous collocations with different hedges had been reported in Old Catalan (cf. Sentí 2017). What is interesting here is that the modal deure with a conjectural value can also appear with a marker expressing different degrees of probability. This is not the case of Italian and French dovere and devoir because according to Squartini (2008) both modal verbs cannot express a conjectural value, only an inferential one (specific and generic inference). He grounds his claim on i) the fact that these verbs cannot co-​occur with epistemic adverbs such as possibly, as shown in (32); and ii) that dovere and devoir do not work when the inference is not based on external evidence (specific inference) or the speaker’s knowledge of the world (generic inference), as in (33): (32)

?

(33)

?

[Suonano alla porta] Forse deve essere il postino. [[The bell rings] Perhaps it must be the postman] [Suonano alla porta] Non aspettava nessuno; deve essere Gianni. [[The bell rings] I was not expecting anybody. It must be Gianni]

Crucially, Catalan deure can co-​occur with different epistemic adverbs, as has been proved above. And it shows the same behavior if we jump to spoken present-​Day Catalan, as in (34) and (35). Moreover, on the contrary to dovere and devoir, in Catalan deure the counterpart of (33) is well-​formed, as (36) shows. (34) són turistes de temporada que tenen una casa aquí […] Ells mateixos… s’ho fan i haha! vénen o no vénen o com vuiguen. I n’hi deu ver molts que també ho lloguen, possiblement. (COD Formentera)

270

Andreu Sentí [they are season tourists who own a house here […] They do everything by themselves and, haha! they come or not come or how they want. And there must be a lot of them who are also renting it, possibly]

(35) També deu sortir més barato, a lo millor. (COD Ciutadella)12 [Also, it must be cheaper, maybe] (36) [Someone rings the bell] No esperava ningú; deu ser la meua veïna. [I was not waiting for anyone; it must be my neighbour]

Finally, another feature of the verb deure in present-​day Catalan shows a conjectural meaning close to uncertainty. The modal deure preserves a fossilized ancient morphological form in the past simple in some Catalan Valencian varieties,13 a tense which has been lost and replaced by a periphrastic past. It seems that this preserved past simple form of deure has specialized in a value of conjecture. As in the previous cases in the 19th c., it also appears in interrogative clauses and can combine with epistemic expressions. Again, it seems that it focuses on the inferential process:

12 This utterance is said by the companion of the informant who is being inter­ viewed for the corpus. 13 In Standard Catalan, in order to build a past form out of deure, the imperfective devia (must.ipfv.3sg) must be resorted to: devia fer (must.ipfv.3sg do). This is because the simple past tense (cantà sing.pst.3sg) has disappeared from the spoken language in most varieties, being replaced by the periphrastic past, formed with the auxiliary anar (‘go’) followed by an infinitive: va cantar (sing.pst.3sg). Since deure cannot enter this periphrastic construction (*va deure fer), the imperfective is used instead: devia fer (must.ipfv.3sg do). An alternative mechanism, the above-​mentioned forms degue/​dega/​deguera, is found in some southern spoken Valencian varieties (Marina Alta, Marina Baixa) (Beltran 2011: 87; Beltran & Herrero 2011: 87; GIEC: § 24.8.4.1). These forms degue/​dega/​deguera are the result of a grammaticalization of the simple past form degué ‘must.pst.3sg’. Crucially, these varieties do not use the past simple any more in any other context or with any other verb in the spoken language –​this situation contrasts with central Valencian, which is exceptional in having maintained the past simple tense for all verbs in the spoken language. To sum up, in the mentioned varieties the forms degue/​dega/​deguera seem to have specialized in a conjectural meaning.

The Modal deure (‘must’) in Contemporary Catalan

271

(37)

La caseta eixa qui la degue fer? (attested,14 Ondara, 2017) [Who must have built that house?]

(38)

[The speaker refers to a new key and the fact of putting it in the key holder]. T'ho degue posar jo [la clau al clauer]. És que no me'n recorde. (attested, Ondara, 2016) [I must have put it there for you. The thing is that I don’t remember]

(39)

[At the university there is a statue of Minerve. The speaker, who is aware of that, talks about a friend who probably saw the statue] Va vore l'estàtua pero no es dega fixar qui era. (attested, Els Poblets, 2016) [She saw the statue but she probably did not pay attention to who it was]

(40)

Veges el retor què deguera dir. (attested, Ondara, 2016) [Imagine what the priest may have said]

All these arguments support the idea that deure is an inferential (evidential) device, even though it can be closely linked to uncertainty. As Cornillie (2009) has argued, the reliability that we can find in each case does not depend on the evidential features of the construction.

5. A Pragmatic Extension: Intersubjectivity Modal constructions usually develop objective, subjective or intersubjective uses. These concepts are different depending on the scholar (Langacker 1987, 1991, 2006; Traugott & Dasher 2002; Nuyts 2001, 2017). However, most agree that there is a tendency toward subjective meanings in diachrony (subjectification). Following Traugott (2003: 128)’s view of intersubjectivity (“speaker/​writter’s attention to the ‘self of the addressee/​reader”),15 14 An attested example refers to utterances produced by native Catalan speakers which have been personally witnessed by the author. 15 Intersubjectivity is understood in different ways by Traugott and Nuyts. Here we are following Traugott (2003)’s definition: “intersubjectivity in my view refers to the way in which natural languages, in their structure and their normal manner of

272

Andreu Sentí

some intersubjective uses can be found in two specific contexts in our data. Firstly, we mean a value of courtesy, illustrated with example (41), which is also present in English must (Mortelmans 2012), as in (42): (41)

Soposo que ja devéu saber lo que vol dir aquesta paraula. (CTILC) [I suppose you must know what that word means]

(42)

“Tell me about the party” “You must have heard about it from my dad. It was Alan and Martha’s twentieth wedding anniversary” (NF, 39; Mortelmans 2012: 2159)

The second intersubjective use of deure is also inferential and it is not found in the Romance counterparts of deure. In this case, the auxiliary deure appears in the future form to express an inference about a future state of affairs. For example, in (43) the speaker says “tu bé deuràs venir demà a caçar” (‘you will likely go out to hunt tomorrow?’). The auxiliary deure in future (deuràs) encodes a reasoning process about a future state of affairs. In other words, a future state of affairs of you going out to hunt tomorrow is presented as the content of an inference.16 (43)

Ah, tú, bè deurás vení demá á cassá? (Ferrer i Codina, Antoni: Lo pagés de l'Ampurdá ó flors trasplantadas, CTILC, 1875) [Ah, you, you will well go out to hunt tomorrow?]

(44)

Lo meu principal va dirme que m’enterés d’alguna agencia de matrimonis; vaig sapigué aquesta, "La Casamentera", y li vaig dí. De modo, que ell deurá vení d'un moment al altre. (Roure, Conrad (Pau Bunyegas): La casamentera, CTILC, 1894) [My boss told me to find out about a marriage agency; I knew about this one, "La Casamentera", and I told him. In such a way that he will come any time now]

operation, provide for the locutionary agent’s expression of his or her awareness of the addressee’s attitudes and beliefs, most especially their ‘face’ or ‘self-​image’ ”. 16 Romance languages cannot express different degrees of certainty in a future state of affairs using future tense. The reason may be the nature of future: it is uncertain by itself (Bybee et al. 1994). In any case, epistemic/​conjectural future must be dis­ tinguished from this use of modal deure inflected for future. Note that using deure in the future form, the state of affairs will be grounded in a subsequent event; while epistemic/​conjectural future refers to a present state of affairs.

The Modal deure (‘must’) in Contemporary Catalan

(45)

273

Bé hi deureu anar, a casa la Roseta? (Iglésias, Ignasi: El cor del poble, CTILC, 1902) [You will well go to Roberta’s house?]

In these examples, the construction “deure +​infinitive” encodes the role of the speaker reasoning, but there is also a dose of intersubjectivity: the speaker wants to focus the attention on the addressee’s attitude. It is as if he said ‘I believe it is reasonable to think that tomorrow you will go out to hunt and I guess you must think the same’. This value is especially found in interrogative sentences which ask the addressee to say something, to confirm the hypothesis. Interestingly, we often find confirmatory question-​tags such as no or oi. This is a pragmatic value that has not been encoded. Therefore, it is not a case of intersubjectification (encoding the intersubjective value) (Traugott 2010), but a case of pragmatic intersubjective values. This pragmatic use is unknown in Old and Modern language, according to our current data, and it is attested at the end of 19th century (corpus CTILC), as seen above, and also in present-​day Catalan (central dialect), in spoken and written texts: (46) Una mica de pa bé ens en deuran portar. (attested, Prats de Lluçanès, 2016) [[Waiting to be served in a restaurant.] Some bread, I suppose they will bring that for us] (47) –​No cal que agafes les ulleres de sol, oi? [per anar a Anglaterra de viatge]. –​ Sortirà poc [el sol], però deurà sortir, no? (attested, Barcelona, 2017) [[Packing to travel to England] –​You don’t need your sunglasses, don’t you? –​ There won’t come out much, but I suppose it will come out, won’t it?] (48)

–​Sí, es nota. Ara deurà anar més de pressa, oi? –​Em penso que sí. (de Pedrolo, Manuel: Mecanoscrit del segon origen: 141; 1974) [–​Yes, I can feel it. Now [the pregnancy belly] I suppose it will grow quicker, won’t it? –​I think so]

(49)

[Un pot amb mostassa] És del Lidl, deurà ser bona. (attested, Barcelona, 2018) [[Talking about a mustard jar recently purchased] It is from Lidl, it must be good.]

274

Andreu Sentí

Although this intersubjective use could be seen as epistemic, we defend that it is a conjectural value (evidential) given that it is oriented to the speaker. In fact, in same tokens, it is obvious that there is no track of epistemicity, as in (50) where the speaker is not downgrading his/​ her epistemic commitment but showing his/​her indifference. Moreover, deure performs a subjectification function as well as it is inflected in future: (50) Ho vaig presentar dimecres al matí i van dir que amb una setmana ens diran alguna cosa. Em fa l’efecte que diran que no. Tant deurà ser… (written text sent through whatsapp, Prats de Lluçanès, 2018) [I submitted it on Wednesday morning and they said that they will answer in a week. I think they will say no. It doesn’t matter]

6. Conclusions The modal construction with deure has undergone a grammaticalization process with a tendency toward subjectification and semantic change from deontic modality to inferential evidential meanings (Sentí 2015a, 2015b). In this paper, we have analyzed the verb deure in the period from 1833 to 1902, using the Contemporary Catalan corpus (CTILC); likewise, we have provided data from spoken present-​day Catalan. At the 19th century, at the beginning of the period of Contemporary Catalan, the core value of the construction changes from deontic necessity to inferential evidentiality. In this regard, we have seen that the discursive tradition is very relevant. Although the deontic meaning is still strong in correspondence, it is fading away in dramatic and narrative texts. Surprisingly, the deontic meaning is still preserved at the beginning of the 20th century. Therefore, the ultimate loss of the deontic meaning is likely situated during the 20th century, because nowadays deure does not preserve this value in most of dialects. Be it as it

The Modal deure (‘must’) in Contemporary Catalan

275

may, the deontic meaning decreased and the inferential one becomes the core meaning. In this article, we have focused on the inferential values of Catalan deure, and we have seen that Squartini’s (2008) inferential gradient (specific, generic inference and conjecture) is useful to account for the diachrony of this verb. The most subjective value (i.e. conjecture) arises during the 19th century. Therefore, this modal verb becomes a more subjective inferential marker, in which the speaker-​oriented feature turns out to be crucial. The speaker plays an important role as a source of evidence. Conjecture is, thus, an extremely subjective value and because of that it seems close to epistemic modality in some contexts. However, we have argued that the modal deure mainains its inferential component even in the most subjective uses or in intersubjective ones. Cross-​ linguistically, the counterparts of deure in Romance (Squartini 2010) and Germanic languages (Mortelmans 2012) do not cover all the inferential gradient nor express the most subjective one, conjecture, with English must being an exception. Differently from all these languages, deure in present-​day Catalan is an inferential marker without any deontic tone in most dialects. The rise of the conjectural meaning and subjective pragmatic uses that we have analyzed may be a clue in the explanation of the loss of the root modal value.

Corpora CIGCMod =​ Corpus Informatitzat de la Gramàtica del Català Modern, dins Martines, J. /​Martines, V. (dirs.) Corpus Informatitzat Multilingüe de Textos Antics i Contemporanis, Alacant, ISIC-​ IVITRA. COD =​ Corpus oral dialectal. [http://​w ww.ub.edu/​cccub/​corp​usor​aldi​ alec​tal-​cod.html] CTILC =​Institut d’Estudis Catalans, Corpus textual informatitzat de la llengua catalana [https://​ctilc.iec.cat/​] Parlars =​ Corpus oral del valencià col·loquial. Universitat de València. [http://​w ww.uv.es/​corv​alc]

276

Andreu Sentí

References Abraham, Werner 2002. Modal verbs. Epistemics in German and English. In Barbiers, Sjef /​Beukema, Frits /​van der Wurff, Wim (eds.) Modality and its interaction with the verbal system, Amsterdam: John Benjamins, 19–​50. Aikhenvald, Alexandra 2004. Evidentiality. Oxford: Oxford University Press. Alemany, Paula 2018. Estudi dels valors modals i evidencials de la perífrasi amb deure al primer terç del segle xx. Degree Final Project, supervised by Andreu Sentí, Universitat de València. Barbet, Cécile 2012. Devoir et pouvoir, des marqueurs modaux ou évidentiels? Langue française 173, 49–​63. Beltran, Vicent 2011. Estudi geolingüístic dels parlars de la Marina Alta. L'empremta mallorquina. Ondara/​ Pedreguer: MACMA/​ IECMA. Beltran, Vicent /​Herrero, Teresa 2011. Estudi geolingüístic dels parlars de la Marina Baixa. L’empremta mallorquina. Ondara/​Pedreguer: MACMA/​IECMA. Bybee, Joan /​Perkins, Revere /​Pagliuca, William 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago/​London: University of Chicago Press. Coates, Jennifer 1983. The Semantics of the Modal Auxiliaries. London: Croom Helm. Cornillie, Bert 2007. Evidentiality and epistemic modality in spanish (semi-​ )auxiliaries: a cognitive-​ functional approach. Berlin: Mouton de Gruyter. Cornillie, Bert 2009. Evidentiality and epistemic modality. On the close relationship between two different categories. Functions of Language 16/​1, 44–​62. Cornillie, Bert /​Marín-​Arrese, Juana /​Wiemer, Björn 2015. Evidentiality and the semantics–​pragmatics interface. An introduction. Belgian Journal of Linguistics 29, 1–​17. Dendale, Patrick 1994. Devoir épistémique, marqueur modal ou évidentiel? Langue française 102, 24–​40.

The Modal deure (‘must’) in Contemporary Catalan

277

Dendale, Patrick 2001. Le future conjectural versus devoir épistémique: differences de valeur et de restrictions d’emploi. Le française moderne 69/​1, 1–​20. Diewald, Gabriele 1999. Die Modalverben im Deutschen. Niemeyer: Tübingen. Diewald, Gabriele 2000. A basic semantic template for lexical and grammaticalized uses of the German modals. Belgian Journal of Linguistics 14, 23–​41. Gavarró, Anna /​Laca, Brenda 2002. Les perífrasis temporals, aspectuals i modals. In. Solà Joan et al. (dir.) Gramàtica del català contemporani, Barcelona: Empúries, vol. iii, 2663–​2726. Goossens, Louis 2003. Patterns of meaning extension, “parallel chaining”, subjectification and modal shifts. In Barcelona, Antonio (ed.) Metaphor and Metonymy at the Crossroads. A Cognitive Perspective, Berlin/​New York: Mouton de Gruyter, 149–​169. de Haan, Ferdinand 1999. Evidentiality and epistemic modality: Setting boundaries. Southwest Journal of Linguistics 18, 83–​101. Kronning, Hans 1996. Modalité, cognition et polysémie: sémantique du verbe modal ‘devoir’. Uppsala/​Stockholm: Almqvist & Wiksell International. Langacker, Ronald W. 1987. Foundations of Cognitive Linguistics, vol. I. Theorical Prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1991. Foundations of Cognitive Linguistics, vol. II. Descriptive Application. Stanford: Stanford University Press. Langacker, Ronald W. 2006. Sujectification, Grammaticization, and Conceptual Archetypes. In Athanasiadou, Angeliki /​Canakis, Costas /​Cornillie, Bert (eds.) Subjectification: Various Paths to Subjectivity, Berlin/​New York: Mounton de Gruyter, 17–​40. Mortelmans, Tanja 2000. On the ‘evidential nature of the ‘epistemic’ use of the German modals müssen and sollen. In van der Auwera, Johan /​Dendale, Patrick (eds.) Modal verbs in germanic and romance languages, Belgian Journal of Linguistics 14, Amsterdam/​Philadelphia: John Benjamins, 131–​148. Mortelmans, Tanja 2012. Epistemic must and its cognates in German and Dutch. The subtle differences. Journal of Pragmatics 44, 2150–​2164.

278

Andreu Sentí

Nuyts, Jan 2001. Epistemic modality, language, and conceptualization: a cognitive-​pragmatic perspective. Amsterdam/​Philadelphia: John Benjamins. Nuyts, Jan 2017. Evidentiality reconsidered. In Marín-​Arrese, Juana Isabel /​Haßler, Gerda /​Carretero, Marta (eds.) Evidentiality revisited: Cognitive grammar, functional and discourse-​pragmatic perspectives, Amsterdam: John Benjamins, 57–​83. Plungian, Vladimir. A. 2001. The place of evidentiality within the universal grammatical space. Journal of Pragmatics 33, 349–​357. Sentí, Andreu 2015a. Modal verbs, future and grammaticalization in Old Catalan. A Cognitive approach. Catalan Journal of Linguistics 14, 179–​198. Sentí, Andreu 2015b. Subjectification and attenuation in the conceptual schema of the Catalan modal verb deure with evidential meaning. eHumanista/​IVITRA 8, 517–​543 Sentí, A. 2017. Modalitat i evidencialitat en català antic. Un acostament cognitiu a les perífrasis verbals amb deure i amb haver. Barcelona: Publicacions de l’Abadia de Montserrat /​IIFV. Sentí, Andreu 2018a. Evidentiality and Epistemic Modality in Old Catalan. A diachronic cognitive approach to the semantics of modal verbs. In Ayoun, Dalila /​Celle, Agnès /​Lansari, Laure (eds.) Tense, Aspect, Modality and Evidentiality: Cross-​linguistic perspectives, Amsterdam/​Philadelphia: John Benjamins, 145–​163. Sentí, Andreu 2018b. El verb modal deure al segle XIX: modalitat deòntica, evidencialitat inferencial, (inter)subjectivitat i mirativitat eHumanista/​IVITRA 14, 646–​672. Sentí, Andreu 2019. Les perífrasis modals catalanes: prescripció i descripció, In Escartí, Josep Vicent (ed). Nunc Dimittis, València: Publicacions de la Universitat de València, 525–​562. Sentí, Andreu 2022. Evidentiality in Catalan. In Wiemer, Björn /​Marín Arresse, Juana Isabel (ed.): Evidential marking in European languages (toward a unitary comparative account), Berlin: De Gruyter (Studia Typologica), 287–​330. Smith, Nicholas 2003. Changes in the modals and semi-​modals of strong obligation and epistemic necessity in recent British English. In Facchinetti, Roberta /​Palmer, Frank /​Krug, Manfred (ed.)

The Modal deure (‘must’) in Contemporary Catalan

279

Modality in Contemporary English, Berlin/​New York: Mouton de Gruyter, 241–​266. Squartini, Mario 2008. Lexical vs. grammatical evidentiality in French and Italian, Linguistics 46:5, 917–​947. Squartini, Mario 2010. Where mood, modality and illocution meet: the morphosyntax of Romance conjectures. In Becker, Martin G. & Remberger, Eva-​Maria (eds.) Modality and Mood in Romance. Modal Interpretation, Mood Selection and Mood Alternation, Berlin/​New York: De Gruyter, 109–​219. Squartini, Mario 2018. Mirative extensions in Romance: evidential or epistemic? In Guentchéva, Zlatka (ed.) Epistemic Modalities and Evidentiality in Cross-​Linguistic Perspective, Berlin/​Boston: De Gruyter Mouton, 196–​216. Traugott, Elizabeth C. 2003. From subjectification to intersubjectification. In Hickey, Raymond (ed.), Motives for Language Change, Cambridge: Cambridge University Press, 124–​139. Traugott, Elizabeth C. 2010. (Inter)subjectivity and (inter)subjectifications: A reassessment. In Davidse, Kristin /​ Vandelanotte, Lieven /​Cuyckens, Hubert (eds.) Subjectification, intersubjectification and grammaticalization, Berlin/​New York, Mouton de Gruyter, 29–​70. Traugott, Elizabeth C. /​Richard B. Dasher 2002. Regularity in Semantic Change. Cambridge: Cambridge University Press. van der Auwera, Johan /​Plungian, Vladimir A. 1998. Modality's semantic map. Linguistic Typology, 2, 79–​124.

Andra Kalnača and Ilze Lokmane

Evidentiality and the Latvian Oblique Forms

Abstract The objective of the paper is to examine the evidential meanings of the oblique forms in Latvian and their interaction with the epistemic and deontic meanings that depend on various syntactic, pragmatic and discourse factors. The oblique forms in Latvian also mark surprise or astonishment, thus evidentiality and epistemicity also convey an overtone of mirativity. In Latvian, oblique forms can also express reportative deontic modality, i.e., reported command and reported necessity. The main meaning of oblique forms is reportative (either quotative or hearsay) thus being evidential. Evidential meaning remains in all oblique uses, regardless of whether discourse mentions or does not mention the source of reported information. Epistemic and mirative connotations are secondary and derived from the core meaning. However, as regards combinations of deontic modality and evidentiality, deontic modal meanings should be regarded as core meanings. Thus, Latvian oblique forms constitute a rather complicated semantic unit in terms of meaning and use. In order to explain this system more clearly, the article takes a closer look at the formation of Latvian oblique forms and their tense paradigm. Keywords: oblique forms, evidentiality, hearsay, epistemicity, mirativity, deontic modality, discourse.

282

Kalnača and Lokmane

1. Introduction The aim of the article is to provide an empirical analysis of the Latvian oblique forms and to look at their evidential functions. Given the diverse use of Latvian oblique forms, it has been important for us, first and foremost, to find out what nuances of evidentiality are characteristic of the Latvian language and which of them should be considered as the core meaning. The article then proceeds to examine which pragmatic and discourse factors determine the expression of evidentiality meanings in Latvian. Second, the article explores the relationship between evidentiality and deontic modality in the use of Latvian oblique forms. As these forms are also used to convey doubt about the truthfulness of the utterance, in addition to evidentiality we will examine their epistemic meanings. Besides, the oblique forms in Latvian also mark surprise or astonishment, thus evidentiality and epistemicity also convey an overtone of mirativity. Epistemicity is not the only modal reading of oblique forms. In Latvian, there is also the possibility of reported commands/​imperatives and necessity; in this latter case evidentiality combines itself with deontic modality. Thus, Latvian oblique forms constitute a rather complicated semantic unit in terms of meaning and use. To explain this system more clearly, the article takes a closer look at the formation of Latvian oblique forms and their tense paradigm. The examples have been taken from different sources: fiction, public media, websites, as well as The Balanced Corpus of Modern Latvian, version 2018 (Līdzsvarotais mūsdienu latviešu valodas tekstu korpuss 2018 (LVK2018), available at: http://​w ww.korp​uss.lv/​id/​LVK2​018). The study does not attempt a statistical analysis of the language units concerned, as this was not its objective. The structure of the article has been developed as follows: first, we outline the theoretical background of the research, then we examine the paradigm of the oblique forms and their core semantics. The next part of the article is dedicated to the empirical analysis of the use of the Latvian oblique forms, with a closer look at contextual meanings and their evidential and epistemic overtones. The article also explores the relationship between evidentiality and deontic modality in the use of Latvian oblique forms.

Latvian Oblique Forms

283

2. Evidentiality and Modality (Theoretical Background) Since the Latvian oblique forms in addition to evidentiality can express modal meanings, the analysis of these forms requires reference to literature that deals not only with evidentiality as such but also with its relationship with modality. There is an enormous volume of literature on the phenomenon of evidentiality and its relationship with modality –​both epistemic and deontic –​and on the notion of modality and the interpretation of its types (among others, van der Auwera/​Plungian 1998; de Haan 1999; Mushin 2001; Aikhenvald 2004; Cornillie 2009; Boye 2012; Portner 2009, 2018). The analysis of the Latvian oblique forms can by no means circumvent the question of the relationship between evidentiality and epistemic modality. According to Portner (2009: 167), ‘[s]‌ome linguists have argued that evidentials and epistemic modals are the same thing. This idea can be phrased in two ways. One can say that epistemic modals are evidentials, or that evidentials are epistemic modals.’ Port­ ner (2009: 264) also states that ‘[e]videntiality should be distinguished from epistemic modality because the two categories relate to quite different types of meaning.’ In our study, we take the Portner’s view that evidentiality and epistemic modality are different categories. However, the use of the Latvian oblique forms suggests that epistemic overtones are also present in the forms carrying evidential meanings. This suggests that evidentiality and epistemic modality are distinct but partly overlapping categories (cf., among others, van der Auwera/​Plungian 1998: 85–​86). While looking at the semantics of the Latvian oblique forms, for the purposes of this paper, we follow a traditionalist approach to evidentiality, in that we distinguish epistemic and deontic modality, without delving into a profound discussion of how these concepts are understood and traditionally described. As a starting point, we take a statement by Aikhenvald (2004: 3) (see also Portner 2018: 247) that “[e]‌videntiality is a linguistic category whose primary meaning is source of information”. As was pointed out by Diewald/​Smirnova (2010: 1), “[t]he basic characteristic of linguis­ tic evidentiality is the explicit encoding of a source of information or

284

Kalnača and Lokmane

knowledge (i.e. evidence) which the speaker claims to have made use of for producing the primary proposition of the utterance”. In a nutshell, evidentiality is a particular semantic category which unites the language means that contain a reference to the source of information and the assessment of the contents of the utterance (Wiemer 2007; 2010: 198; see also Holvoet 2001: 111). Epistemic modality points out the text author’s attitude toward the contents of the utterance (judgement, assertion) which can involve a need, possibility, or probability (Palmer 2001: 8–​9). Deontic modality connotes the speaker’s degree of requirement of, desire for, or commitment to the realization of the proposition expressed by the utterance (Palmer 2001: 9–​10). The structure of the article is based on its theoretical background. First, Section 3 describes the core semantics of the oblique forms and the system of their forms. Section 4 then deals with the contextual use of the oblique forms. Sections 5 and 6 deal more closely with the relationship between the oblique forms and epistemic and deontic ­modalities.

3. Core Semantics and Paradigm Latvian is one of the relatively few languages that have developed specific forms for the expression of evidentiality –​the so-​called oblique mood (e.g. Plungian 2001, 2010; Aikhenvald 2004: 33; see also Keha­­­­ yov 2008a and Lavine 2010). Oblique forms are used to indicate that the author of a text is not the source of the information contained in that text (e.g. Endzelīns 1951: 902; Holvoet 2007: 81–​82; Kalnača 2014: 122; Kalnača/​Lokmane 2021: 240). Oblique is used whenever the speaker wants to refer to an utterance of another person, i.e., to express reportative (or reportive) evidentiality: (1)

Viens one.nom.sg

no   kandidātiem    premjera of   candidate.dat.pl  prime_​ minister.gen.sg

285

Latvian Oblique Forms biroja vadītāja   postenim office.gen.sg head.gen.s  position.dat.sg esot bijušais veselības ministrs. be.obl.prs fomer.ptcp.nom.sg health.gen.sg minister.nom.sg [Reportedly, one of the candidates for the position of the head of the prime minister’s office is the ex-​minister for health.] (LVK2018)

Example (1) states that the speaker has obtained information from someone else, although no specific source has been identified. Thus the oblique form is undoubtedly an evidential, as it can be the only marker of the source of information, and evidentials have the indication of evidence as their core meaning rather than pragmatic inference. Although verbum dicendi is not compulsory, oblique can be used in indirect speech sentences to emphasize that information was obtained from what others have said (example (2)), or even to indicate a specific source of information (example (3)). In this case, the evidential marking is embedded, namely, it appears in the subordinate clause: (2)

Man I.dat

sacīja, told.pst.3

ka that

pret against

sauli sun.acc.sg

derot be_​useful.obl.prs

arī pūderis. also powder.nom.sg [I was told that powder could also do as a sun-​screen.] (Z. Skujiņš) (3)

Sieviete woman.nom.sg

noprasīja, ask.pst.3

kurš who.nom.sg

no from

bērnības childhood.gen.sg

atceroties kādu skaitāmpantiņu. remember.obl.prs some.acc.sg nursery_​rhyme.acc.sg [The woman asked who could remember a nursery rhyme from their childhood.] (LVK2018)

286

Kalnača and Lokmane

In example (2) the verbum dicendi in the main clause sacīja ‘told’ indicates that there will be a narrative follow-​up –​someone who told the author about his or her experience of protecting the skin from the sun that is too strong and bright. The oblique form derot in the subordinate clause indicates that the author of the text is not the originator of this information. In example (3), however, the main clause specifically refers to the source of information (the woman). Although traditional Latvian grammars treat oblique forms as forming a distinct verbal mood (so called atstāstījuma izteiksme, also relatīvais mods –​among others, Endzelīns 1951: 902, 976–​983; Nītiņa/​Grigorjevs 2013: 483–​486, 495–​496; Kalnača 2014: 122–​126), there is still much dis­ cussion about the exact nature of the relationship between evidentiality and modality. That the problem of the relation between evidentiality and epistemic modality also concerns Latvian is witnessed by the fact that oblique forms are described in Latvian grammars in the overall framework of other moods thus addressing the possibilities of their modal use (i.e. modal use of the oblique forms) along with non-​modal reportative meanings (e.g. Nītiņa/​Grigorjevs 2013: 495–​496; Kalnača/​Lokmane 2021: 240–​244). The fact that the oblique forms in Latvian can have epistemic overtones clearly shows that evidentiality and epistemicity are interrelated (for a theoretical discussion see, e.g. de Haan 1999; Mushin 2001; Kehayov 2008b; Cornillie 2009; Boye 2012). As we will see in Section 6 of this article, oblique forms can also be used to express deontic modality. This fact, too, proves that evidentiality and modality are interrelated. In our paper, we shall not delve into the classification of the oblique forms in Latvian grammars, as this is a field of research on its own right. As regards the semantics of the oblique forms, we believe that the core meaning is reportative. The remaining meanings, namely, epistemic connotations, are secondary and derived from the core meaning. However, as regards combinations of deontic modality and evidentiality, deontic modal meanings should be regarded as core meanings (see Section 6 for more detailed discussion). In Latvian, oblique forms are marked by means of the suffix -​ot (non-​reflexive verbs) or -​oties (reflexive verbs), see examples (1)–​(2) and (3)–​(4), expressing either present or future tense meanings depending on the kind of verbal stem (present or future) they are added to (e.g. Nītiņa/​ Grigorjevs 2013: 483–​484; Kalnača 2013: 84; 2014: 123–​124; Kalnača/​

287

Latvian Oblique Forms

Lokmane 2021: 240–​244). They can be synthetic or analytical (periphrastic). The compound oblique forms are formed by combining auxiliary es-​ot ‘reportedly am/​is/​are’ vai būš-​ot ‘reportedly will be’ with the past active participle (op. cit.), see examples (5)–​(6) and Table 1. Since the oblique mood lacks person forms, person meanings are derived from clause subjects. Table 1.  The paradigm of oblique forms (adapted from Kalnača 2013: 84). Tense

Synthetic form

Analytical (periphrastic) form

PRS

las-​ot ‘I, you, he/​she, etc. reportedly read’ lasīš-​ot ‘I, you, he/​she, etc. reportedly will read’

es-​ot lasījis,-​usi, -​uši, -​ušas ‘I, you, he, she etc. reportedly had read’ būš-​ot lasījis,-​usi, -​uši, -​ušas ‘I, you, he, she, etc. reportedly will have read’

FUT

(4)

Esmu dzirdējusi, ka daļa publikas par tām latviešu mīlas lietām [izrādes iestudējumā] arī apvainojas, –​ ka jūs d[aktieri] smej- ​oties that you.nom.pl [actors] pok_​f un-​obl.prs par vienkāršo cilvēku dzīvi. ptcl ordinary.gen.pl people.gen.pl life.acc.sg [I have heard that some of the audience gets upset about these Latvian love things [at the show] –​that you [the actors] are reportedly poking fun at the lives of ordinary people.] (LVK2018)

(5)

Jau already

vēstīts, announce.ptcp.nom.sg

ka that

deputāts deputy.nom.sg

es-​ot be.aux-​obl.prs

sarīkojis stirr_​up.ptcp.nom.sg

dzērumā drunkenness.loc.sg

nekārtības slimnīcā. brawl.acc.pl hospital.loc.sg [It has already been announced that the MP had reportedly stirred up a brawl at hospital.] (Kas Jauns)

288 (6)

Kalnača and Lokmane

Viņš he.nom

samaksāš-​ot repay-​obl.fut

pēc after

tam, this.dat.sg

kad when

būš-​ot be.aux-​obl.fut

parādu debt.acc.sg

saņēmis receive.ptcp.nom.m

plānoto naudu. plan.ptcp.acc.f money.acc.f [He says he will repay the debt when, as he says, he will have received the expected amount of money.] (LVK2018)

The auxiliary verb esot ‘reportedly am/​is/​are’ is often omitted in the present perfect oblique in longer stretches of text, in tales, fairy tales, etc., especially if the verbum dicendi is given in a context (e.g. Endzelīns 1951: 977; Nītiņa/​Grigorjevs 2013: 484; Kalnača 2014: 124–​ 125). Thus, in example (7) a narrated fragment of the text is introduced by the main clause, in which there is a verbum dicendi in the form of the past perfect bija stāstījis ‘had told’, which indicates that the following ancient events in the utterance are retold by grandfather’s father. As is typical in the Latvian language, this narrative uses an oblique present perfect form with an omitted auxiliary esot ‘reportedly am/​is/​are’: (7) Par deviņpadsmito gadu vectēvam daudz bija stāstījis viņa tēvs, kas lielinieku laiku kopā ar Morbergu bija pārlaidis namā iepretim Bastejkalnam. Kādu nakti pie durvīm pavēloši [esot] klauvēts. Morbergs [esot] uzvilcis un atkal [esot] novilcis ziemas mēteli ar lapsādas oderi un [esot] atdevis Ulstem. Jaunās varas mērķis tomēr [esot] izrādījies cits –​Morberga dzīvokļa un mantas atsavināšana. Paglābtais mētelis Morbergam vēlāk [esot] labi noderējis. [Grandfather had heard a lot about the year nineteen nineteen from his father, who had spent the Bolshevik period together with Morbergs in a house opposite Bastejkalns (downtown Riga). One night they heard someone knocking demandingly at the door. Morbergs put his fox-​f ur collar winter coat on then took it off again and gave over to Ulste. It turned out though that the new government had other intentions –​alienation of Morberg’s apartment and belongings. The rescued coat proved useful to Morbergs later.] (Z. Skujiņš)

289

Latvian Oblique Forms

However, if the oblique form is absent from the context, then a phrase without an auxiliary verb does not allow one to ascertain whether the phrase in question is in the oblique or indicative mood (for a more detailed treatment of this question, see Muižniece/​Metslang/​ Pajusalu 1999; Kalnača 2000; Ivulāne 2015: 86–​87; Kalviša 2018: 65–​ 67). (8)

Vairāki several.nom.pl

avoti source nom.pl

uzskata, believe.prs.3

ka   šī      cilvēka    loma that   that.gen.sg  man.gen.sg  role.nom.sg valsts state.gen.sg

politikas politics.gen.sg

veidošanā shaping.loc.sg

kādu some.acc.sg

[ir/​esot]

bijusi be.ptcp.nom.sg

laiku time.acc.sg

daudz   lielāka      nekā  premjerministram. much    greater.nom.sg  than  Prime_​ Minister.dat.sg [Several sources believe that the role of that man in shaping state politics for some time has (reportedly) been much greater than that of the Prime Minister.] (LVK2018)

Both in examples (7) and (8) there are oblique forms with an omitted auxiliary esot ‘reportedly am/​is/​are’ –​the reportedness is signaled by a verbum dicendi or a verb of thinking in the main clause –​ bija stāstījis ‘had told’ (7) and uzskata ‘believe’ (8). Cases in which indirect speech situations can be conveyed by both oblique and indicative forms will be discussed in Section 5 of this article.

290

Kalnača and Lokmane

4. The Context of Use The terminology used while describing evidentiality, the notions of reportative, quotative, hearsay, citation, echoic and other uses, is widely discussed in the theoretical literature (see, among others, Korotkova 2017; Holvoet 2018; Mañoso-​Pacheco/​Juárez-​Escribano 2019). On the other hand, in terms of sentence structure, there is an ongoing discussion about the relationship of evidentiality with reported speech (direct and indirect speech) as well as the primary proposition expressed in the sentence (AnderBois 2014; Chojnicka 2012, 2016; Mañoso-​Pacheco/​ Juárez-​Escribano 2019). In this study, ‘reportative’ is used as a superordinate notion to designate information that comes from someone else’s verbal report and thus refers to all subdomains of reported evidentials. The term ‘quotative’ is used to refer to ‘a reported evidential which includes an exact indication of who provided the information’ (Aikhenvald 2004: 25). ‘Hearsay’, on the contrary, contains no reference to the author, time, and circumstances of the utterance (Holvoet 2018: 6), thus the exact source of the reported utterance is not specified. Reported speech, in its turn, ‘brings together tools and devices used for attributing knowledge to another speaker’ (Chojnicka 2012: 173) and is usually divided into direct and indirect speech. Direct speech gives a verbatim quotation of what somebody else has said and uses quotation marks, while in indirect speech the quoted text is rendered by a subordinate clause. In both cases, the source of the information is indicated in a separate proposition. It is acknowledged that the notion of reportative evidentiality often overlaps with that of reported speech, however, ‘reportive evidentiality focuses more on coding of somebody else’s information and the statement that “there is some evidence” rather than the source of the reported assertion itself’ (Mañoso-​Pacheco/​Juárez-​Escribano 2019: 10, see also AnderBois 2014; Korotkova 2017). In our view, the understand­­ ing of reported speech primarily follows from the syntactic structure of the sentence and the particularities of the organization of the discourse, while reportative evidentiality denotes the fact that information had been derived from what others have said. Consequently, reported

291

Latvian Oblique Forms

speech and reportative evidentiality should not be considered as one type of phenomenon or component of a single cline (as stated in Cho­ jnicka 2012) but as two different dimensions that interact in discourse. Thus, it is not surprising that Latvian uses oblique mainly in indirect speech utterances. Therefore, we will move on to discuss the use of oblique forms in different contexts and we will describe the relationship between the context and the sentence structure with the types of reportative evidentiality. 4.1. Oblique as Quotative in Indirect Speech Constructions The oblique forms are often used when transforming direct speech into indirect speech even if there is no doubt about the veracity of the contents. In such cases, verbum dicendi is used in the main clause (thus speech report is a complement of the speech verb), and the source of information is specified (among others, Chojnicka 2008, 2016; Kalnača 2011: 290; Nītiņa/​Grigorjevs 2013: 496). (9)

ielas street.gen.sg

satiku rakstnieku meet.pst.1sg writer.acc.sg

kurš who.nom.sg

teica, tell.pst.3

ka that

Latvijas Latvia.gen.sg

Enciklopēdijā Encyclopedia.loc.sg

Reiz once

uz on

vajagot redaktorus. need.obl.prs editor.acc.pl [Once I bumped into a writer on the street who told me that the Latvian Encyclopedia needed editors.] (Ieva)

Unlike the phrases with the direct speech that explicitly quote what someone says, the perspective of the author of the utterance has changed slightly, and he reports what the writer he has just meet says from his point of view, selecting only the part of information he considers relevant.

292

Kalnača and Lokmane

Some authors believe that solely statements that do not specify a specific source of information are truly evidential (e.g. Chojnicka 2012: 178–​179); in other cases, the same formal marker should be considered as a reported speech marker. It has also been stressed that reportative evidentials do not form a separate proposition (among others, Holvoet 2018; Mañoso-​Pacheco/​Juárez-​Escribano 2019: 12). Since in the above example the source of the information is stated in the main clause, thus expressed in a separate proposition, one might consider that the oblique form merely duplicates the content already stated and is not a prototypical evidential. However, we believe that the evidential nature of the oblique form does not change and that a partial repetition of information in discourse is a common phenomenon. Such duplication could rather be attributed to the presence of various epistemic overtones, which will be further discussed in Section 5. 4.2. Oblique as Hearsay in Indirect Speech Constructions The oblique mood can also be used in such indirect speech phrases where the subjectless main clause contains a verbum dicendi as a predicate. Thus, the source of the information is unknown, or it stems from an anonymous author, or, in other words, the judgment is based on something that has been said (10) or is often said (11) by unspecified persons: (10)

Toties ptcl

runā, say.prs.3

ka that

visā whole.loc.sg

Latvijā Latvia.loc.sg

viens lācis tomēr esot. one.nom.sg bear.nom.sg ptcl be.obl.prs [Instead, it is rumored that there indeed is one bear in the whole of Latvia.] (LVK2018)

293

Latvian Oblique Forms

(11)

Kuluāros corridor.loc.pl

baumo, rumor.prs.3

ka that

rakstnieks writer.nom.sg

dzejoļu poem.gen.pl

rakstīšanu writing.acc.sg

esot atmetis        un be.aux.obl.prs abandon.ptcp.nom.sg  and pievērsies    hroniku sacerēšanai move_​on.ptcp.nom.sg  chronicle.gen.pl writing.dat.sg par latviešu literatūru on Latvian.gen.pl literature.acc.sg 21. gadsimtā. 21st century.loc.sg [It is rumored in corridors that the writer has abandoned his poems and moved on to writing chronicles on 21st-​century Latvian literature.] (www.­ satori.lv)

If sentences with a specific author’s reference could be classified as second-​hand information, then this case should be classified as third-​ hand information. Such oblique narrative is often used by journalists to mitigate their responsibility for what they say, but at the same time, they indirectly shape public opinion based on rumor, probability, and speculation. The less explicit are the references to the source of information, the more salient becomes the opinion of the author of the text. It should be noted that the oblique mood is not used to refer to common knowledge. 4.3. Oblique as Quotative –​the Source of Information Mentioned in the Context Cases where the source of information is mentioned in the context or at least understood from the context could also be considered as quotative expressions, as it is clear who is responsible for what is said (Palmer 2001, 73–​74):

294 (12)

Kalnača and Lokmane

Meitas daughter.nom.pl

viņai she.dat.sg

par about

Jāni Jānis.acc.sg

sāka start.pst.3

stāstīt. tell.inf

Cik how

viņš he.nom.sg

esot be.cop.obl.prs

labs! good.nom.sg

Kā how

viņš he.nom.sg

savu own.acc.sg

māti mother.acc.sg

apkopjot! take_​care.obl.prs [The daughters started telling her about Jānis. How good he is! How well he takes care of his mother!] (LVK2018)

In example (12), it is clear from the context that the narrators are the daughters, while example (13) in terms of content is like indirect speech, but the structure of the first sentence does not allow the speech verb stāsta ‘tell’ to directly add a complement clause: (13) Kā nonākusi līdz šādiem apstākļiem, viņa stāsta neskaidri. Zaudējusi darbu, lose.ptcp.nom.sg job.acc.sg [sieviete]

trīskārtējas maiņas three_​time.gen.sg change.gen.sg

rezultātā result.loc.sg

esot be.aux.obl.prs

bez without

dzīvokļa, flat.gen.sg

tad than

vienu one acc.sg

[esot]

palikusi remain.ptcp.nom.sg

brīdi    it kā time acc.sg  as if

dzīvojusi patversmē. live.ptcp.nom.sg social_ ​home.loc.sg [How she got to that point, she does not tell clearly. Lost her job, [the woman] changed her flat three times and, as a result, lost it, then for some time reportedly stayed at a social home.] (Diena)

295

Latvian Oblique Forms

4.4.  Oblique as Hearsay –​the Source of Information is not Mentioned In this case, there is no context with verbum dicendi. In such instances oblique it is the only marker showing that the information is obtained from someone else’s verbal report, as utterance is obviously presupposed, but it is not clear who is responsible for the utterance. (14) Laiks weather.nom.sg

būšot be.cop.obl.fut

labs. good.nom.sg

Nu, nu, redzēsim. interj interj see.fut.1pl [The weather would be good. Well, well, let’s see.] (LVK2018)

Thus, in Latvian, oblique in the function of hearsay can be used in the text in two ways: in indirect speech constructions, as well as in a separate simple clause where there is no indication as to the author of the presupposed utterance at all. Though, in many cases this can be predicted depending on the subject being discussed or inferred from a wider discourse, for instance, in example (14), the reported information is most likely provided by the meteorologists. In addition to the four contextual uses of the oblique forms discussed here, there is a specific syntactic construction including oblique forms and closely linked to certain epistemic and even mirative meanings, which will be briefly examined in Section 5.

5. Evidentiality and Epistemicity Evidential markers can develop epistemic overtones. As Holvoet (2018) states, ‘[i]‌t is well known that the use of an evidential marker may be a device allowing the speaker to distance her/​himself from a claim and thus indirectly to express an epistemic stance’. Epistemic overtones occur mainly when the content of the reported utterance is likely to

296

Kalnača and Lokmane

be questionable. Thus, a speaker can use the oblique mood to express doubts about the truthfulness of the narrated content, to voice his or her disbelief about it, and to distance him-​or herself from any credibility assessment. 5.1. Epistemic Agnosticism The use of verbum dicendi in the main clause often implies that the oblique form in the sub-​clause(s) is not compulsory. In terms of meaning, in many cases, there is not a great difference, whether the indicative or oblique is used, and information is re-​narrated without any doubt about the veracity of content. Someone else has told the speaker the information in question, but s/​he has no evidence as to its truth value. In this case, the utterance contains no epistemic overtones at all. The fact that in some cases there is no semantic difference between the indicative and the oblique is shown by example (15), where the auxiliary sentence uses both moods: (15)

Zemniece farmer.nom.sg

atzīst, admit prs.3

ka that

sausuma drought.gen.sg

no from

var can.ind.prs.3

glābt save.inf

laistīšanas    sistēma, watering.gen.sg  system.nom.sg taču tas esot ļoti dārgi. but that.nom.sg be.cop.obl.prs very costly [The farmer admits that an irrigation system could save her from a drought but that would be very costly.] (www.tvnet.lv)

However, if the author of the utterance is completely sure of the contents, for example, if it relates to a fact that is generally known or verifiable, the indicative is preferred, as in example (16):

297

Latvian Oblique Forms

(16)

Satversmē Constitution.loc.sg

teikts, provide.ptcp.nom.sg

ka that

Valsts State.gen.sg

prezidentam President.dat.sg

ir be.cop.ind.prs.3

tiesības right.nom.pl

ierosināt propose.inf

Saeimas atlaišanu. Saeima.gen.sg dissolution.acc.sg [The Constitution provides that the President of the State is entitled to propose the dissolution of the Saeima.] (LVK2018)

5.2. Doubt Therefore, the choice of the oblique form in indirect speech may be a signal of certain doubt about the veracity of the utterance, often combined with critical examination and irony. This mainly depends on discourse factors –​shared knowledge, type of discourse, the topic discussed, etc. For instance, weather forecasts are often not taken at their face value, especially if the weather forecast refers to a more remote future, because such forecasts often fail to be true. (17)

Šogad this_​year

būšot be.obl.fut

ļoti very

silta warm.nom.sg

vasara, prognozē sinoptiķi. summer.nom.sg say.prs.3 meteorologist.nom.pl [This year a very warm summer is to be expected, the meteorologists say.] (Kas Jauns)

Journalists quite often choose to use the oblique form in the title and in the text itself, thus expressing skepticism and a negative epistemic stance:

298

Kalnača and Lokmane

(18) Valsts state.nom.sg

gribot want.obl.prs

pati itself.nom.sg

izputināt banku? destroy.inf bank.acc.sg [The state itself would admittedly try to ruin the bank?] (www.tvnet.lv)

The author of a text (utterance) can add an epistemic overtone to an evidential phrase (re-​narrated information) by means of the particles it kā ‘as if’, laikam ‘probably’ etc., thus doubting the veracity of the reported information. The meaning of doubt is enhanced but the oblique still retains its function as an evidential marker. Such parallel use of the oblique and of the modal particle is not surprising, because evidential markers and epistemic modals evoke the same conceptual base: (19)

Limbažos Limbaži.loc.pl

pilsētas town.gen.sg

svētkos festival.loc.pl

it kā būšot rokgrupas koncerts. as if be.obl.fut rock-​band.gen.sg cert.nom.sg Tādas baumas es šodien dzirdēju!!! [In Limbaži town festival, a rock-​band is likely to perform. Such rumors I heard today!!!] (LVK2018)

The author of a text can also convey not only doubt but also a negative attitude toward the reported information by means of an epistemic particle: (20) Es I.nom

it kā as if

neesot not_​be.cop.obl.prs

pieņemama acceptable.ptcp.nom.sg

Briselei Brussels.dat.sg

savas own.gen.sg

nostājas view.gen.sg

striktās strong.gen.sg

dēļ of

par Krievijas jautājumu. on Russia.gen.sg issue.acc.sg [Presumably, I am not acceptable to Brussels because of my strong views on the Russian issue.] (LVK2018)

299

Latvian Oblique Forms

The meanings of doubt and affective evaluation are often reinforced by the so-​called contrastive context (Chojnicka 2016). In exam­ ple (21), the first sentence contains an oblique form, while the next sentence expresses a contrasting view: (21) Man I.dat

nupat just

teica, tell.pst.3

ka that

es I.nom

esot be.aux.obl.prs

lībiešu Livonian.gen.pl

sabiedrības society.gen.sg

aizgājusi. leave.ptcp.nom.sg

Bet but

tad then

es I.nom

domāju, think.pst.3

kas what

tā this.nom.sg

lībiešu Livonian.gen.pl

sabiedrība society.nom.sg

no from

vispār tāda ir? ptcl such.nom.sg be.prs.3 [I was just told that I had, as it were, left Livonian society. But then again, I ask myself: what is it, this Livonian society?] (LVK2018)

5.3. Disbelief Evidential markers may develop an epistemic overtone ‘contrary to what somebody has said’ (see Holvoet 2018 about Lithuanian hearsay marker esą ‘as if’). In this case, a negative opinion of another person or their utterance is expressed, namely, the content of the reported information is clearly not to be believed. The negation of the content is usually due to pragmatic factors –​to the context (22) or to general knowledge, such as the fact that the much predicted end of the world does not come (23) or that modern society tends to reject extreme (national, leftist, etc.) views (24):

300 (22)

Kalnača and Lokmane

Lielākā biggest.nom.sg

anekdote par anecdote.nom.sg about

mani I.acc

es I.nom

esot be.aux.obl.prs

ieguvis acquire.ptcp.nom.sg

izglītību education.acc.sg

un and

tāpēc therefore

ir –​ be.prs.3

kļuvis nepieejams. become.ptcp.nom.sg unapproachable.ptcp.nom.sg [The best joke about me is this: I acquired education and that is why I became unapproachable.] (LVK2018)

(23)

Pareģo kārtējo pasaules galu –​ prophesy.prs.3 another.acc.sg world.gen.sg end.acc.sg tas būšot klāt jau rīt. it.nom.sg be.obl.fut present already tomorrow [Yet another end of the world is being prophesied, it would, they say, arrive tomorrow.] (www.tvnet.lv)

(24)

Es esot par daudz nacionāls, I.nom be.cop.obl.prs too much national.nom.sg man esot kreisi uzskati u. c. I.dat be.obl.prs left.nom.pl belief.nom.pl [I am, as it were, too nationalistic, I am told to be a leftist, and so forth.] (LVK2018)

The quantitative aspect, too, testifies to the influence of pragmatic factors on the epistemic overtones of the oblique mood. Chojnicka (2008: 57) has found that (concerning Latvian): ‘Very often the oblique forms are (...) found in speculative texts. Obviously, the less safe, less clear is the topic discussed or described, the more oblique forms are used. (...) Most oblique forms are found in speculative writings (0.182 forms per sentence). These are followed by comments (0.146), informative (0.132) and, finally, analytical texts (0.114 forms per sentence).’

Latvian Oblique Forms

301

The fact that the oblique forms in Latvian can have epistemic overtones clearly shows that Latvian is not different from other languages in that evidentiality and epistemicity are interrelated (for a theoretical discussion see, e.g. de Haan 1999; Mushin 2001; Cornillie 2009; Boye 2012). 5.4. Defeated Expectation The polysemantic periphery of evidentiality includes exclamative utterances that are linked to mirativity and epistemicity (on mirativity and its interaction with evidential and modal meanings see, mentioning all but a few, Lazard 1999; DeLancey 2001; Aikhenvald 2012; Rett/​ Murray 2013; Simeonova 2015). As Aikhenvald (2004: 9) states, ‘[a]‌n unexpected evidentiality choice may acquire additional stylistic overtones of sarcasm, irony, or indignation.’ This is also the case in Latvian (e.g. Kalvišaa 2018: 62–​63). In such utterances, the verb būt ‘be’ most often appears in the oblique form of the simple present, i.e., esot, as a copula to a nominal predicate, while the function of the subject is performed by the demonstrative pronoun tas ‘this /​that’ ((25)–​(26)), tāds ‘such’ (27) that would relate to some contextual item referred to earlier, thus forming a particular syntactic construction. (25)

Kā var bārstīties ar apvainojumiem, kurus nevar pierādīt? Kas tie tādi ir par policijas darbiniekiem? Un tie esot and this.nom.pl be.cop.obl.prs profesionāļi!!! professional.nom.pl [How can one spread insinuations that cannot be proved? Who are those police officers? These are called “professionals”!?] (www.pietiek.com)

302 (26)

Kalnača and Lokmane

Un and

grīda! floor.nom.sg

Tas this.nom.sg

esot be.cop.obl.prs

parkets! parquet_​floor.nom.sg

Melns kā darva! [And the floor. Is this called “parquet floor”? Pitch black, it is.] (A. Eglītis) (27) Fui, phew

ne no

kauna, shame.gen.sg

ne no

goda, honour.gen.sg

un tāds esot cilvēks? and such.nom.sg be.cop.obl.prs human.nom.sg [Phew, no shame, no honor, would you call this a human being?] (www.­ dzirkstele.lv)

Such sentences express defeated expectation of sorts, with added negative evaluation of that fact. The author of the utterance had hoped for a different result, basing him-​or herself on a generally accepted knowledge of a situation or fact, i.e., an assumption that police officers are expected to be professionals (25), or that the inlaid floor is well maintained which has not been the case (26), while the example (27) shows the author’s reaction to a web-​page comment that he or she considers impolite. Therefore, one can consider that evidentiality in such cases has been partially abstracted, thus the meaning of mirativity and epistemic modality here becomes more salient. In Latvian, we can speak of three competing meanings –​the evidential one (that is partly faded), epistemic (disbelief, irony) and mirative (unexpected, mostly unpleasant, event). However, the relationship between mirativity and evidentiality in the use of Latvian oblique forms are a subject of further research, and in this paper they will not be discussed in more detail.

303

Latvian Oblique Forms

6. Evidentiality and Deontic Modality In Latvian, the oblique forms are used not only to report imaginary, fantastic and unbelievable events but also in the cases when one needs to report a command, demand or necessity expressed by a person (or persons). Therefore, it can be claimed that in Latvian one is able to express reportative deontic modality; furthermore, reported command and reported necessity show morphological and syntactic differences. 6.1. Reported Imperative A specific use of oblique together with the subordinating conjunction (or particle) lai combines evidential and imperative (deontic) meanings (Kalnača 2011: 290, 295; Kalviša 2018: 63–​64; Kalnača/​Lokmane 2021: 259). This usually occurs in complex sentences, where the verbum dicendi is in the main clause, while the subordinate clause has the predicate in an oblique form (usually in the simple present): (28) Un viņi and they.nom

[radinieki] [relatives]

[man] [me]

saka, tell.prs.3

lai sub

es I.nom

nākot come.obl.prs

[dzīvot viņu mājā], [to live in their house]

viņi they.nom

man I.dat

došot give.obl.fut

istabiņu augšstāvā. room.acc.sg upper_​floor.loc.sg [And they [relatives] tell [me] to come [to live in their house], they would give me a small room at the upper floor.] (www.dzivesstasts.lv)

A sentence can also have a more complex syntactic structure. In example (29), the verbum dicendi that indicates the fact of reporting

304

Kalnača and Lokmane

is in the subordinate clause, which in turn is subordinated to another subordinate clause with the reported command: (29) Kad mājas saimnieks teicis, lai [kaimiņi] savācot [izbērtās smiltis], sub [neighbour] clear.obl.prs [scatter sand] [tie] teikuši: jā, bet nevākuši. [they] say.ptcp.nom.pl yes but not_​clear.ptcp.nom.pl [When the house owner told [the neighbours] to clear [the scattered sand] away, [they reportedly] said: yes, but didn’t.] (Latvijas Avīze)

In such cases we can speak of the second-​hand imperative (see, e.g. Aikhenvald 2004: 250–​253). Here, too, there is a particular syntactic construction –​subordinate clause of the object introduced by the subordinating conjunction lai, while the oblique form is in the simple present tense (other oblique forms are not used in such cases). The fact that the subordinate clause contains an initial imperative is demonstrated by the conjunction lai, which is also used in the function of the particle adjacent to the 3rd person imperative form of the verb (SG/​PL) (Nītiņa/​G rigorjevs 2013: 491; Kalnača/​Lokmane 2021: 254–​255): (30) Lai ptcl

[viņš] [he]

atnāk, come.imp.3

es ar viņu parunāšu! I.nom to he.ins.sg talk.fut.1sg [Let [him] come, I’ll talk to him!] (LVK2018)

It can be claimed that in the event of the reported imperative featuring a subordinate clause, i.e., introduced by lai, it concurrently functions as both particle and subordinating conjunction. In addition, there are cases (example (31)) where, in the same complex sentence, lai functions both as a subordinate conjunction and expresses a purpose (lai atbrīvotu), and combines the function of the subordinate conjunction and the deontic particle with the oblique form (lai mēs ejot):

305

Latvian Oblique Forms

(31)

Meitene, kura [hokeja arēnā] ierāda vietas, tā vietā, lai atbrīvotu mums mūsu conj liberate.cond we.dat we.gen norādītās indicated.acc.pl

biļetēs ticket.loc.pl

vietas, seat.acc.pl

teica, lai mēs ejot citur. say.pst.3 sub we.nom go.obl.prs somewhere_​else [The usherette [at the hockey arena], instead of liberating the seats we had as on our tickets, said we should go somewhere else.] (www.sportacentrs.com)

In the case of the imperative, the verbum dicendi is in the main clause, but could also be replaced by a verb that exhorts, commands, etc., for example, (pa)māt ‘to wave’. This suggests that such a sentence is probably a fragment of some more extensive reported event, cf. example (32) and see also below example (33): (32)

Un bez jebkādas gaidīšanas vārtsargs man pamāja ar roku, lai es ejot cauri [vārtiem]. conj I.nom go.obl.prs through [gate] [And, without any delay, the goalkeeper waved me by the hand, inviting to pass through the [gate].] (LVK2018)

There may also be cases where the verbum dicendi is not formally expressed but from the context a conversation has occurred where one participant has repeatedly asked or advised another. In reporting such a situation (also indicated by the oblique forms of renarration), lai is used in conjunction with an oblique form, in addition to reproducing the structure of the spoken utterances –​dividing compound phrases into separate textual units, as in the following example: (33) Kantora darbiniece pār brillēm nopētīja viņu [Agru] kā jukušu. Nolika malā pusrakstīto kvīti. Kad nauda atradīšoties, tad pabeigšot rakstīt. [..] Lai aizņemoties [naudu]. No radiem vai draugiem. Viņa saprotot, tagad visiem esot spiedīgi apstākļi. Lai atnākot vēlāk. Viņi šovakar strādājot ilgi. [..] Lai ejot uz sociālās nodrošināšanas nodaļu, tur iedošot pabalstu, kantoriste vēl nosauca pakaļ.

306

Kalnača and Lokmane [An employee of the office cast a glimpse over her glasses at her [Agra] as if dealing with a lunatic. She put down a half-​w ritten receipt. When the money is found, she said, she would then finish the writing. [..] Let her borrow [some money]. From relatives or friends. She understands that now everyone has hard times. Let her come back later. They are working long hours tonight. [..] Let her go to the Social Security Department, she would get a benefit there, the employee of the office shouted, as she was leaving.] (LVK2018)

As the introduction of the fragment mentions an employee of an office, it is her who recommended that the narrator of the text, Agra, should set off her debt and come back later to continue the conversation. 6.2. Reported Necessity Necessity in Latvian is expressed by a special verbal mood, the debitive. The debitive mood is used to name a necessary, obligatory action, i.e., an action that must be performed (among others, Skujiņa 2007: 416; Nītiņa/​Grigorjevs 2013: 497; Kalnača 2014: 128). The form of the debitive is formed with the prefix jā-​ and the auxiliary verb būt in a finite (present, past, or future) form, and the experiencer of the necessity is in the dative (in detail see, e.g. Holvoet 2007: 173; Kalnača 2014: 128–​130; Lokmane/​Kalnača 2014): (34)

a. Es I.nom

zinu, ka know.prs.1sg that

ir jā-​maksā, be.aux.prs.3 deb-​pay un es maksāšu. and I.nom pay.fut.1sg [I know I must pay, and I will.] (LVK2018)

man I.dat

307

Latvian Oblique Forms

b. Tagad sāksies sējas laiks, un man būs jā-​pērk minerālmēsli. and I.dat be.aux.fut.3 deb-buy fertilizer.nom.pl [Now the sowing time will begin, and I will have to buy fertilizer.] (LVK2018)

The meaning of the forms of the debitive is usually associated with deontic modality. In Latvian, deontic modality is mainly manifested in the imperative mood, which is used to express the speaker’s assessment of an action to influence the addressee to fulfil the speaker’s intention –​to make sure that the action is performed. However, deontic modality can also be communicated by means of the debitive mood, as in example (34). In that case, however, an addressee is not addressed directly. Since the debtor is in the dative case, the speaker can only describe an action as being obligatory for him/​herself or others. Necessity in Latvian could also be reported by means of grammatical forms, where the form of the debitive is used in conjunction with an auxiliary verb in the oblique form –​both in the present tense (35) and in the future tense (36). Latvian grammars usually refer to these forms as the oblique subtype of the debitive mood (vajadzības izteiksmes atstāstījuma paveids –​e.g. Nītiņa, Grigorjevs 2013: 488, 497), although the existence of such forms calls into question the status of both the oblique mood and the debitive mood in Latvian, according to Holvoet (2007: 177–​178). (35)

Jautāts, kā vislabāk nokļūt templī, ceļotājs stāsta, ka ērti ir noīrēt rikšu ar vadītāju. Līdz Ankorvata templim to Angkor_​Wat.gen.sg temple.dat.sg esot be.aux.obl.prs

jā-​brauc deb-​d rive

30–​40 minūtes. 30–​40 minute.nom.pl [When asked how best to arrive at the temple, the traveler says that it is convenient to charter a rikshaw with the driver. Apparently, it is a 30–​40 minute drive to the Angkor Wat temple.] (www.delfi.lv)

308

Kalnača and Lokmane

(36) Kārlis smējās, ka pagrabs būšot jā-​paplašina. that cellar.nom.sg be.aux.obl.fut deb-​expand [Kārlis laughed and said that the cellar needed to be expanded.] (LVK2018)

As can be seen from examples (35)–​(36), which also contain reported debitive, the sentences (usually in the main clause) normally contain a verbum dicendi. In addition, one can also find sentences that combine the so-​called ordinary oblique form and the debitive form of the auxiliary verb in the oblique form: (37) Vecāki parent.nom.pl

teica, tell.pst.3 galda nesēdīšoties, table.gen.sg not_​sit.obl.fut

ka that

bez manis without I.gen

pie at

jo because

pirmā first.nom.sg

īstā real.nom.sg

nodarbību classes.gen

diena day.nom.sg

fakultātē esot jā-​atzīmē. faculty.loc.sg be.aux.obl.prs deb-​celebrate [My parents told me that they would not sit at the table without me because the first day of classes at the faculty should be celebrated.] (G. Priede)

In the case of reported necessity, the deontic meaning of necessity is attributable to the reported proposition. Thus, in such cases, it is possible to speak of a combination of evidentiality and deontic modality (Lokmane/​Kalnača 2014: 181–​182; Kalviša 2018: 71–​72). However, the fact that, in Latvian, a reported imperative is also possible, and that the debitive form may be supplemented by an auxiliary oblique form, indicates that the evidential and deontic meanings in this case are not mutually dependent, but are rather realized in parallel. Of course, in such cases one might ask, which meaning –​the deontic or the evidential –​is central. However, the fact that forms with

Latvian Oblique Forms

309

deontic meaning can be used in a reportative context clearly indicates that deontic meaning is the core meaning.

7. Concluding Remarks The oblique forms in Latvian always have to do with an external speech act. Their main meaning is reportative (either quotative or hearsay) thus being evidential. Evidential meaning remains in all oblique uses, regardless of whether discourse mentions or does not mention the source of reported information. Epistemic and mirative meanings are superimposed, as it were, on the evidential ones –​the evidential meaning serves as a basis to which the author adds additional epistemic connotations or overtones of counter-​expectation and negative surprise. Epistemic overtones of the oblique forms range from neutral or agnostic to doubt and disbelief, and largely depend on pragmatic factors –​the context and shared assumptions. The epistemic and mirative meanings of the oblique forms are most manifest in the presence of additional modality indicators, for instance, modal particles or special syntactic constructions. The modal semantics of these devices would push evidentiality further to the background, but it would be still present in the given utterance. In the case of reported imperative and reported necessity evidentiality combines with deontic modality. Such a combination takes place mainly in complex sentences with verba dicendi in the main clause, and the meaning of the deontic command or necessity refers to the reported proposition.

310

Kalnača and Lokmane

Acknowledgments This study was carried out in the framework of the Latvian National Research Program “Letonika –​Fostering a Latvian and European Society”, project “Research on Modern Latvian Language and Development of Language Technology” [grant number VPP-​LETONIKA-​2021/​ 1– ​0006].

Abbreviations 1, 2, 3 person acc accusative aux auxiliary cond conditional conj conjunction cop copula dat dative deb debitive fut future gen genitive ind indicative inf infinitive ins instrumental interj interjection loc locative LVK2018 The Balanced Corpus of Modern Latvian 2018 nom nominative obl oblique ptcl particle ptcp participle pl plural prs present pst past

Latvian Oblique Forms

311

ptcp participle sg singular sub subordinator

Sources Diena (newspaper) Eglītis, Anšavs 1993. Misters Sorrijs. Rīga: Liesma. Ieva (weekly) Kas Jauns (weekly) Latvijas Avīze (newspaper) Līdzsvarotais mūsdienu latviešu valodas tekstu korpuss 2018. [The Balanced Corpus of Modern Latvian, version 2018.] Available at: http://hdl.handle.net/20.500.12574/11 Priede, Gunārs 2016. Saniknotā slieka. In Struka, Ieva (ed) Gunāra Priedes dzīve un darbi. 1977–​1986, Rīga: Jumava, 355–​393. Skujiņš, Zigmunds 1999. Miesaskrāsas domino. Rīga: Preses nams. www.delfi.lv (news portal) www.dzi​rkst​ele.lv (municipal news portal) www.dzi​vesz​ina.lv (portal of miscellaneous interests) www.piet​iek.com (news portal) www.sat​ori.lv (internet magazine) www.sport​acen​trs.com (news portal) www.tvnet.lv (news portal)

References Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: Oxford University Press. Aikhenvald, Alexandra Y. 2012. The essence of mirativity, Linguistic Typology 16, 435–​485, https://doi.org/10.1515/lity-2012-0017.

312

Kalnača and Lokmane

AnderBois, Scott 2014. On the exceptional status of reportative evidentials, Semantics and Linguistic Theory 24, 234–​254. https://​doi. org/​10.3765/​salt.v24i0.2424. Boye, Kasper 2012. Epistemic Meaning. A Crosslinguistic and Functional-​Cognitive Study. Berlin/​Boston: De Gruyter Mouton, https://​doi.org/​10.1515/​978311​0219​036. Chojnicka, Joanna 2008. Atstāstījums latviešu preses rakstos par politiku. [Reporting in Latvian Press Articles on Politics.] MA thesis. Poznań: Adam Mickiewicz University in Poznań. Chojnicka, Joanna 2012. Reportive evidentiality and reported speech: is there a boundary? In Usonienė, Aurelija /​Nau, Nicole /​ Dabašinskienė, Ineta (eds) Multiple Perspectives in Linguistic Research on Baltic Languages, Newcastle upon Tyne: Cambridge Scholars Publishing, 170–​192. Chojnicka, Joanna 2016. Latvian verbs of speaking and their relations to evidentiality. Kalbotyra 69, 59–​83. https://​doi.org/​10.15388/​ Klbt.2016.10367. Cornillie, Bert 2009. Evidentiality and epistemic modality: on the close relationship of two different categories. Functions of Language 16/​1, 44–​62, https://​doi.org/​10.1075/​fol.16.1.04cor. DeLancey, Scott 2001. The mirative and evidentiality. Journal of Pragmatics 33, 369–​ 382. https://​doi.org/​10.1016/​S0378-​ 2166(01)80001-​1. Diewald, Gabriele /​Smirnova, Elena 2010. Introduction. Evidentiality in European languages: the lexical-​grammatical distinction. In Diewald, Gabriele /​Smirnova, Elena (eds) Linguistic Realization of Evidentiality in European Languages, Berlin/​New York: De Gruyter Mouton, 1–​25. https://​doi.org/​10.1515/​978311​0223​972. Endzelīns, Jānis. 1951. Latviešu valodas gramatika. [A Grammar of Latvian.] Rīga: Latvijas Valsts izdevniecība. de Haan, Ferdinand 1999. Evidentiality and epistemic modality: setting boundaries. Southwest Journal of Linguistics 18, 83–​101. Holvoet, Axel 2001. Studies in the Latvian Verb. Kraków: Wydawnictwo Uniwersytetu Jagiellońskiego. Holvoet, Axel 2007. Mood and Modality in Baltic. Kraków: Wydawnictwo Uniwersytetu Jagiellońskiego.

Latvian Oblique Forms

313

Holvoet, Axel 2018. Epistemic modality, evidentiality, quotativity and echoic use. In Guentchéva, Zlatka (ed) Epistemic Modalities and Evidentiality in Cross-​Linguistic Perspective, Berlin: De Gruyter Mouton, 242–​258. Ivulāne, Baiba 2015. Palīgnozīmē lietotu darbības vārdu sistēma latviešu valodā. [System of Latvian Verbs Used as Auxiliaries.] PhD thesis. Rīga: Latvijas Universitāte. Kalnača, Andra 2000. Izteiksmju sinonīmija latviešu valodā, [Synonymy of moods in Latvian.] Baltu Filoloģija IX, 173–​186. Kalnača, Andra 2011. Atstāstījuma izteiksme un modalitāte palīgteikumos ar lai, [Relative mood and modality in subordinate clauses with lai] Baltistica XLVI/​2, 286–​298. Kalnača, Andra 2013. Morfoloģija. In Veisbergs, Andrejs (ed) Latviešu valoda, [Morphology. The Latvian Language.] Rīga: LU Akadēmiskais apgāds, 45–​108. Kalnača, Andra 2014. A Typological Perspective on Latvian Grammar. Warsaw/​Berlin: De Gruyter Open. https://​doi.org/​10.2478/​ 978311​0411​317. Kalnača, Andra /​Lokmane, Ilze 2021. Latvian Grammar. Rīga: University of Latvia Press. https://​doi.org/​10.22364/​latg​ram.2021. Kalviša, Liene 2018. Evidencialitāte latviešu valodā. [Evidentiality in Latvian.] Rīga: LU Akadēmiskais apgāds, https://doi. org/10.22364/lr.5. Kehayov, Petar 2008a. An Areal-​Typological Perspective to Evidentiality: The Cases of the Balkan and Baltic Linguistic Areas. PhD thesis. Tartu: Tartu University Press. Kehayov, Petar 2008b. Interactions between grammatical evidentials and lexical markers of epistemicity and evidentiality: a case-​study of Bulgarian and Estonian. In Plungian, Vladimir /​Wiemer, Björn (eds) Lexikalische Evidenzialitäts-​Marker in slavischen Sprachen. Wiener Slawistischer Almanach, Sonderband 72, München: Sagner, 165–​201. Korotkova, Natasha 2017. Evidentials and (relayed) speech acts: hearsay as quotation. Proceedings of SALT 25, 676–​694. http://​dx.doi. org/​10.3765/​salt.v25i0.3969.

314

Kalnača and Lokmane

Lavine, James E. 2010. Mood and transitivity restriction in Lithuanian: the case of inferential evidential. Baltic Linguistics 1, 115–​ 142, https://​doi.org/​10.32798/​bl.437. Lazard, Gilbert 1999. Mirativity, evidentiality, mediativity, or other? Linguistic Typology 3, 91–​ 109. https://​doi.org/​10.1515/​ lity.1999.3.1.91. Lokmane, Ilze /​Kalnača, Andra 2014. Modal semantics and morphosyntax of the Latvian DEBITIVE. In Leiss, Elizabeth /​Abraham, Werner (eds) Modes of Modality. Modality, Typology, and Universal Grammar, Amsterdam/​Philadelphia: John Benjamins Publishing Company, 167–​192. https://​doi.org/​10.1075/​slcs.149. Mañoso-​Pacheco, Lidia /​Juárez-​Escribano, Beatriz 2019. Reportive evidentiality. A theoretical revision. Revista Inclusiones 6/​4, 189–​206. Muižniece, Liena /​Metslang, Helle /​Pajusalu, Karl 1999. Past participle finitization in Estonian and Latvian. In Erelt, Martin (ed) Typological Studies. III, Tartu: University of Tartu, 128–​157. Mushin, Ilana 2001. Evidentiality and Epistemological Stance: Narrative Retelling. Amsterdam: John Benjamins. https://​doi.org/​ 10.1075/​pbns.87. Nītiņa, Daina /​Grigorjevs, Juris (eds) 2013. Latviešu valodas gramatika. [A Grammar of Latvian.] Rīga: Latvijas Universitātes Latviešu valodas institūts. Palmer, Frank R. 2001. Mood and Modality. Cambridge: Cambridge University Press. Plungian, Vladimir A. 2001. The place of evidentiality within the universal grammatical space. Journal of Pragmatics 33/​3, 349–​357. https://​doi.org/​10.1016/​S0378-​2166(00)00006-​0. Plungian, Vladimir A. 2010. Types of verbal evidentiality marking: an overview. In Diewald, Gabriele /​ Smirnova, Elena (eds) Linguistic Realization of Evidentiality in European Languages, Berlin/​ New York: De Gruyter Mouton, 15–​58. https://​doi.org/​10.1515/​ 978311​0223​972. Portner, Paul 2009. Modality. Oxford: Oxford University Press. Portner, Paul 2018. Mood. Oxford: Oxford University Press.

Latvian Oblique Forms

315

Rett, Jessica /​Murray, Sarah E. 2013. A semantic account of mirative evidentials. Proceedings of SALT 23, 453–​472. http://​dx.doi.org/​ 10.3765/​salt.v23i0.2687. Simeonova, Vesela 2015. On the semantics of mirativity, Proceedings of the 2015 Annual Conference of the Canadian Linguistic Association. Ottawa: University of Ottawa; available at: http://​cla-​acl. ca/​w p-​cont​ent/​uplo​ads/​Simeon​ova-​2015.pdf. Skujiņa, Valentīna (ed) 2007. Valodniecības pamatterminu skaidrojošā vārdnīca [The Dictionary of Basic Linguistic Terms]. Rīga: Latvijas Universitātes Latviešu valodas institūts. van der Auwera, Johan /​Plungian, Vladimir A. 1998. Modality’s semantic map. Linguistic Typology 2, 79–​124. Wiemer, Björn 2007. Lexical markers of evidentiality in Lithuanian. Italian Journal of Linguistics 19, 173–​208. Wiemer, Björn 2010. Hearsay in European languages: toward an integrative account of grammatical and lexical markers. In In Diewald, Gabriele /​Smirnova, Elena (eds) Linguistic Realization of Evidentiality in European Languages, Berlin/​New York: De Gruyter Mouton, 59–​130. https://​doi.org/​10.1515/​978311​0223​972.

Section C Evidentiality and Modality in Discourses and Genres

Natalia Mora-​López

Evidential Expressions in Spanish Accounts of Religious Miracles of the 17th Century

Abstract Evidential expressions or linguistic realizations of the information on which a statement is based (Aikhenvald, 2004) are an important resource when dealing with the veracity of a text, since depending on the way it is related to evidence, the text may seem closer to or further from the truth. A type of text where conveying veracity to the message may play a key role is that in which the narrated events may look unbelievable, unnatural or even supernatural like accounts of religious miracles which were commonly reported and widely accepted some centuries ago, in the 16th and 17th centuries in Spain (García de Enterría, 1994, 1996, 1998). In order to enrich and widen the study of evi­­ dential expressions in Spanish, this paper examines their role in three Spanish accounts of religious miracles of the 17th century (Sánchez Iglesias, 2017) and tries to shed some light on the way the author pres­ ents the information as true by means of evidential strategies. Results show the texts draw on reported evidence as well as verbs of seeing and hearing to support their narratives: people say something (typically quoted), people see something (e.g. miracles and signs), and people hear something (e.g. saints’ voices). Keywords: 17th century, evidentiality, hearing, miracle accounts, sight, Spanish.

320

Natalia Mora-López

1. Introduction In the last decades, there has been an increasing interest in the phenomenon of linguistic evidentiality in the fields of pragmatics and semantics (Dendale and Tasmowski 2001; Aikhenvald 2004). Strictly speaking, evidentiality refers to “a linguistic category whose primary meaning is source of information” (Aikhenvald 2004: 3), that is, a reference to the way in which knowledge was obtained (Ferrari and Gallardo 1999; Laca 2000; Palmer 2001). There has been some debate about the limits of evidential expressions due to their overlap with other meanings such as epistemicity (de Haan 1999; Cornillie 2009), which refers to the “evaluation of the chances that a certain hypothetical state of affairs under consideration (or some aspect of it) will occur, is occurring or has occurred in a possible world” (Nuyts 2001: 21). Recent approaches, like Aikhenvald’s, recognize evidentiality as a linguistic category in itself, whereas some more traditional analyses (Bybee 1985; Palmer 2001; McCready and Ogata 2007) see evidentials as a category inside epistemic modality. However, according to the definition above, evidentiality is the linguistic realization of the information source on which a statement is based (Aikhenvald 2004) and, therefore, evidential expressions may not include any information about speaker’s certainty regarding the statement or whether it is true or not, since they just supply the source of information. As Comrie (2000: 1) points out, evidentials started to be widely studied in the mid to late nineteen eighties (e.g. Givón 1982; Chafe and Nichols 1986; Willet 1988), but they have only recently been addressed in Spanish, although with a very successful reception and spread (Marín Arrese, Hidalgo and Molina 2002, 2004; Bermúdez Wacht­­­ meister 2004; Cornillie 2007, 2008, 2015; Marín Arrese and Carretero 2014; Ahern, Amenós-​Pons and Guijarro-​Fuentes 2015; Carretero and Zamorano-​Mansilla 2015; Cornillie and Gras Manzano 2015; Vatri­­ can 2015). As can be expected, evidential expressions are an important resource when dealing with the (intended) veracity of a text, since depending on the way it is related to evidence, the text may seem closer

Evidential Expressions in Spanish Miracle Accounts

321

to or further from the truth. A special type of text where conveying veracity to the message may play a key role is that in which the narrated events may, at first, look unbelievable, unnatural or even supernatural. Such texts include accounts of religious miracles which, although rarely seen nowadays, were commonly reported and widely accepted some centuries ago, in the 16th and 17th centuries in Spain (García de Enterría 1994, 1996, 1998). Thus, in order to enrich and widen the study of evidential expressions in Spanish as well as shedding some light on the way authors present information as truly as possible in apparently adverse contexts, this paper examines the role of evidential markers and expressions oriented to reliability issues in three Spanish accounts of religious miracles of the 17th century (Sánchez Iglesias 2017). The remainder of the paper is organized as follows: section 2 presents the hypothesis and research questions on which the study is built; section 3 summarises the main types of evidence and evidential strategies used in Spanish; section 4 describes the corpus used for the study; section 5 analyses the most relevant occurrences of evidentiality in the corpus as well as those related to reliability and the senses of sight and hearing; section 6 presents the quantitative results regarding the distribution of occurrences; finally, section 7 draws some conclusions from the analysis.

2. Hypothesis and Research Questions This study explores the hypothesis that authors turn to specific strategies to provide evidence and reliability when writing accounts of a controversial incident like miracles. More specifically, it tries to answer the following research questions: • • • •

Do authors use evidential markers when reporting miracles? If they do, what types of evidence do authors use? How are they related to reliability? Are there any types of evidence that are common to several accounts, or are they author-​specific?

322

Natalia Mora-López

3. Types of Evidence, Spanish Evidentiality, and Reliability Before examining the texts, it is important to establish the theoretical framework used for the analysis. This involves determining a general categorization of the evidence and a more specific type of elements to be looked for. 3.1. Types of Evidence In addition to the expressions used to introduce evidence in discourse, it is important to set a division of the types of evidence they may refer to. An elaborate account of evidentiality can be found in Aikhenvald (2004), although for the purposes of this study, the simpler system established by Willett (1988) is comprehensive enough. Moreover, Willett’s system focuses on a specific subfield of evidentiality, namely reported speech, which is more closely related to the contents that are found in the corpus analyzed here. Willett distinguishes between two types of evidentiality, namely direct and indirect evidentiality (Willett 1988: 57). These two types as well as their inner divisions are shown in Figure 1. Direct

Attested

Indirect

Reported

Inferring

Visual Auditory Other sensory Second-​hand Third-​hand Folklore Results Reasoning

hearsay

Figure 1.  Types of evidence (Willett, 1988: 7)

Direct evidentiality is found when the speaker indicates a first-​ hand, sensory evidence for the transmitted information. On the other hand, indirect evidentiality is in turn divided into two types. The first

Evidential Expressions in Spanish Miracle Accounts

323

one, reported, covers second-​and third-​hand evidence (i.e. the speaker received the information from either a direct witness or not a direct witness, respectively) as well as information belonging to established oral history or folklore. The second one, inferring evidence, is given when the assertion made by the speaker is marked to have been deduced or inferred, focusing on either the results obtained or the reasoning ­followed. 3.2. Evidentiality and Evidential Strategies in Spanish As mentioned above, evidentiality refers to the source of the information presented. It may be realized in language by means of a number of strategies, including morphological, lexical or syntactical. Since the purpose of this paper focuses on specific evidential realizations in accounts of religious miracles and not on the discussion of the notion of evidentiality itself, a simple classification will suffice as working notions. A summary of evidential expressions in Spanish is presented in Table 1, based on the descriptions made by López Ferrero (2005) and González Ruiz, Izquierdo Alegría and Loureda Lamas (2016). Table 1.  Summary of Spanish evidential expressions Adverbs and adverbial locutions

Nouns

Adjectives

al parecer [it seems], por lo visto [so it seems], según parece [as it seems] aparentemente [apparently], claramente [clearly], evidentemente [evidently], visiblemente [visibly], obviamente [obvi ously], personalmente [personally], presumiblemente [presumably], presuntamente [allegedly], supuestamente [supposedly], virtualmente [virtually], etc. consecuencia [consequence], conclusión [conclusion], evidencia [evidence], inducción [inducement], inferencia [inference], obviedad [obviousness], etc. aparente [apparent], claro [clear], evidente [evidente], indiscutible [indisputable], obvio [obvious], presumible [presumable], presunto [alleged], supuesto [supposed], etc. (Continued)

324

Natalia Mora-López

Figure 4.  Continued Verbs

Communicative verbs: advertir [warn], afirmar [claim], anunciar [announce], apuntar [point], comunicar [communicate], dar a entender [lead to believe], decir [say], definir [define], escribir [write], explicar [explain], exponer [expose], indicar [indicate], manifestar [express], mencionar [mention], notificar [notify], precisar [specify], proclamar [proclaim], reseñar [describe], revelar [reveal], señalar [point out], subrayar [underline], etc. Inductive verbs: concluir [conclude], demostrar [prove], hallar [find], implicar [imply], inducir [induce], inferir [infer], obtener [obtain], suponer [suppose], etc. Perceptive verbs: aparecer [appear], apreciar [appreciate], distinguir [distinguish], encontrar [find], observar [observe], parecer [seem], presentar [present], reflejar [reflect], ver [see], etc. (Semi)auxiliary verbs followed by infinitive: amenazar [threaten], deber de [must], prometer [promise], resultar [result], tener que [have to], etc.

Verb tenses

present perfect, future simple, going to, imperfective past, conditional. se ve [it looks], veo [I see], ves [you see], ya veo [I see], ya ves [you see]

Expressions based on the verb ver Discourse en todo caso [in any case], a ver [let’s see] markers Prepositional phrases introduced by según [according to] Structures Que viene a cenar [that he is coming to have dinner] introduced by a quotative que [that]

It can be seen that evidential expressions are not limited to verbs at all, but other word classes may be involved too, including nouns, adjectives, adverbs, prepositional phrases and discourse markers, or even fixed expressions and clausal structures. However, like in English (c.f. Chafe, 1986: 262), evidentiality in Spanish is most often, but not only, expressed by lexical means instead of grammatical, in contrast to

Evidential Expressions in Spanish Miracle Accounts

325

other non-​European languages (see Chafe and Nichols, 1986; Willett, 1988; Palmer, 2001; Aikhenvald and Dixon, 2003; Aikhenvald, 2004). Due to the exploratory nature of this study on the use of evidentials in accounts of miracles, the combination of Willett’s (1988) more general classification of evidentiality and a more specific account of the realizations it may have in Spanish, following López Ferrero (2005) and González Ruiz, Izquierdo Alegría and Loureda Lamas (2016), is suitable enough to provide a working frame for the analysis. 3.3. Evidentiality and Reliability Although a more detailed description of the corpus is presented in section 4, it is important to mention here that authorship at the time when the texts were written was not an important issue. For this reason, the author’s involvement and presence in the text may be scarce or even inexistent. The author tries to disappear to show or describe the facts as something that just “happened”. However, the fact that they narrated these facts and wrote about them at the time may be seen as another evidence or validation of their existence: the author does not bring the events as a controversial, open-​to-​debate incident, and no signs are shown of the author disbelieving or distancing from the facts. Therefore, authors do not necessarily mention how they came to know about the facts they narrate, but they describe them as if they were personally watching and hearing everything the protagonists say, including quotations of the exact words uttered, transcripts of letters (and therefore, a written evidence) and the like. For this reason, a broader analysis is done to include not only evidential expressions used in relation to events but also verbs and other items related to seeing and hearing things as well as reliability expressions used to emphasize the truthfulness of the information transmitted. This last group may overlap with the concept of epistemic modality, that is, those occurrences addressing the estimation of the chances of a proposition to be true or false (Nuyts 2001; Carretero and Zamorano-​Mansilla 2013). A thorough analysis of epistemic elements, although interesting, was beyond the scope of this study. However, a set

326

Natalia Mora-López

of expressions used to prove the reliability of the accounts when referring to certifying information is discussed in section 5.6, since they typically include a specific reference to whom certifies them and have been regarded as evidence of the events taking place. Therefore, only those expressions of reliability dealing with the source of information have been included in the analysis.

4. Corpus The analysis of evidential expressions in Spanish accounts of religious miracles is based on three texts belonging to a wider corpus collected by Sánchez Iglesias (2017). This larger corpus consists of a catalogue of 400 accounts printed in Spain, either in Madrid or Seville, in the 17th century, currently located in the Library of the Royal Academy of History in Madrid. They are part of the emergence of journalism in Spain and they combine history, journalism and literature; therefore, their main purpose was to inform citizens about a variety of topics such as weddings, battles, festivities, tournaments, etc. In this sense, these accounts were also informing of the occurrence of miracles. They were mainly transmitted orally, since most of the population could not read. Nine out of those 400 accounts refer to religious miracles. They all share the same structure: title, heading (short statement about the content) and body (the narrative itself), and they are usually four pages long. The three texts (8802 words) studied here were randomly selected from those nine. The first text (abbreviated here as “Sim”) is titled Relación de la colocación que se hizo al reverendíssimo padre maestro fray Simón de Roxas a dos días del mes de febrero deste presente año de 1625 [Account of the placement done to the very revered father Simón de Roxas on February 2nd, 1625]. It consists of 1834 words and is a compilation of miracles related to the priest Simón de Roxas. In the second text (“Moy”), Relación de un prodigioso milagro que el apóstol de Oriente san Francisco Xavier hizo con una religiosa de el convento de la Purificación de la villa de Muymenta [Account of

Evidential Expressions in Spanish Miracle Accounts

327

the prodigious miracle that the apostle from East st. Francisco Xavier performed with a nun from the Purification convent in Muymenta village], a nun suffers from a very serious illness, from which she miraculously recovers thanks to the image of a saint; however, she knows she will die in six months even though she feels healthy, and she eventually dies when she said she would. It consists of 3429 words. Finally, the third text (“Mun”), Relación de algunos de los muchos milagros que a obrado Dios nuestro Señor en Munébrega, lugar de la comunidad de Calatayud, por medio de una imagen de San Ignacio de Loyola, fundador de la Compañía de Iesús [Account of some of the many miracles that God our Lord worked in Munébrega, in the community of Calatayud, through a statue of st. Ignacio of Loyola, founder of the Society of Jesus], is a compilation of short reports of many miracles produced by the use of a saint’s statue and oil and consists of 3539 words. Due to the limited textual sample (only three texts) and consequent number of evidential occurrences, this is mainly a qualitative study regarding what type of evidential expressions are found in the texts. However, similar results could be expected in other samples. Quantitative results (section 6) are obtained regarding the distribution of the different types of occurrences found.

5. Results and Discussion Three texts about accounts of religious miracles were analyzed to find evidential expressions and strategies that provided the source from which the special type of information presented was obtained. Additionally, some other strategies used to infuse the text with a higher level of reliability were also taken into account. In general, it must be pointed out that a modest amount of standard evidential expressions in the author’s voice were found in the texts, which indicates a specific style in reporting information at this time in this genre. The absence of these expressions aims to leave the author’s viewpoint out of the narrative, since this figure is not important

328

Natalia Mora-López

in this period. Thus, the texts look more objective and factual by putting together a series of events that happen one after the other, such as conversations (either rephrased or accurately transcribed with the protagonists’ words as they uttered them) and actions taking place as a consequence of someone seeing or hearing something. Therefore, in this type of texts, the result obtained is that the author tries to disappear to give the reader the impression that the story comes directly from the protagonists as they perceived it (as they saw, heard and said things), with a consequent profuse use of expressions related to seeing and hearing, as the basis for the evidence, and verbs of saying. To analyze evidential expressions, they have been divided into several categories: three types of indirect evidence (inferential, reported speech and folklore) and one type of direct evidence (attested by the author) as well as the use of references to sight and hearing, and expressions related to reliability, due to the importance and presence they have in the texts and their relevance for evidentiality. 5.1. Indirect Evidence: Inferential Inferential evidence is obtained by means of deduction and similar processes triggered by the perception of an event. In these texts, most of the cases of indirect inferential evidence found are related to the verb parecer [seem] and sight-​related nouns, adverbs and verbs, as shown in the examples below. (1)

parece nos basta para que la tengamos por misteriosa (Moy) [It seems that it is enough for us to have it as mysterious]

In (1), the text reflects the author’s voice and judgment or inference when saying that the circumstances that surrounded the event seem to be enough to regard it as mysterious. Additionally, the author refers to the first-​person plural to refer to himself, so it could be paraphrased as It seems to be enough for me, thus introducing himself in the text twice: once explicitly by using the evidential parece [seems],

Evidential Expressions in Spanish Miracle Accounts

329

and secondly by alluding to himself with the pronoun nos [us]. Let us now consider (2). (2)

Aquí digo que fue para oýr y ver aquella dichosa alma, sintiéndose regalada con la villa de aquel su serafín, que parece vino a Moymenta en busca suya para por la mano la llevar de la tierra al cielo. (Moy) [Here I say that he went to hear and see that happy soul, who felt gifted with the company of that seraph, who seems to have come to Moymenta in search of her to take her from earth to haven by the hand]

Here, when the author talks about the visit of the saint, he says que parece [seems to] have come to find her and take her to heaven. It can be thought that when talking about saints or God’s intentions or interpreting their actions, this author is a bit more cautious in his language choices. It is also important to note that the author refers to himself in the first person singular, which is very unusual and explicitly introduces his own words. Actually, it may even be interpreted that when he writes Aquí digo [Here I say], he makes reference to his own opinion, since it could be paraphrased as “this is something that, in my opinion, is remarkable and worth mentioning”. Finally, this example also includes a reference to sight and hearing ( fue para oýr y ver [he went to hear and see]). A slightly different style of evidential expressions is found in (3), where the author introduces the word muestra [sign, proof], which is typically related to sight. In fact, the author draws on sight as a resource to emphasize how crowded the church was (nunca se ha visto tan ocupada [it has never been seen so crowded]). (3)

concurrió tanta gente en la iglesia que nunca se ha visto tan ocupada, muestra bastante de que su aclamación es la de Dios. (Sim) [so many people attending the church that it had never been seen so crowded, proof enough of the fact that his words are God’s]

Some other inferential evidentials are not in the author’s voice, but in the witnesses’. Since one of the most common resources in these texts is transferring the narrative power to the protagonists’ voices

330

Natalia Mora-López

themselves, evidentials can be found in their speech too. This is the case of (4). (4)

pero llegando a mi celda, llevada parece de Dios que no quiso que la gloria del santo Xavier se pudiesse atribuir a nuestro padre… (Moy) [but reaching my room, guided apparently by God who did not want that the saint Xavier’s glory could be attributed to our father…]

This is part of one of the copies of the letters in the report of the nun’s miracle. She is talking about her experience, and she says it was God who was apparently guiding her and her actions. Again, when talking about a supernatural entity’s intention, the words used are more cautious. 5.2. Indirect Evidence: Reported Speech This category includes those examples where the author transfers the narrative power to the witness of the events by introducing the exact words uttered through direct speech or a rephrasing with indirect speech. This information may come either from an oral or a written source. Thus, reporting verbs (decir [say], responder [reply] confesar [confess], afirmar [claim], etc.) will have an extensive presence in the texts. As mentioned above, the author mostly transmits the protagonists’ own words. This is promoted in the text itself and perceived as the most accurate report of the events. Actually, in one of the texts, the author writes (5). (5)

Y porque lo más milagroso del sucesso no se puede exprimir mejor que con las palabras de quien en sí le esperimentó, pondremos aquí fielmente la copia de dos cartas de la enfermedad escritas al mismo padre Baltasar Saravia, en la primera de las quales dize assí: (Moy) [And because the most miraculous side of the event cannot be expressed better than with the words of who experienced it himself, we will faithfully present here the copy of the two letters on the illness written to father Baltasar Saravia himself, in the first of which he says:]

Evidential Expressions in Spanish Miracle Accounts

331

The author prefers to transfer his authority to the witnesses and protagonists of the accounts, and he explicitly says so. After that, a transcription of the letter is inserted in the text. The use of direct speech is abundant. Some examples are presented below in (6)-​(9). (6)

Un niño que estava con garrotillo, ya en el último trance de la muerte, con voz debilitada dixo: (Sim) [A boy who suffered from dipthteria, in the last moments before passing away, said with a weak voice:]

(7)

Ella respondió: “…”. En esto llegó el padre Valerio Piquer y dixo: (Mun) [She replied: “…”. Then, father Valerio Piquer arrived and said:]

(8)

A esto respondió ella estas formales palabras: (Moy) [To this she replied these formal words:]

(9)

…entró el padre Roxas y dixo: (Sim) […father Roxas came in and said:]

Actually, as will be seen at the end of this section, the most profuse resource in these texts is direct speech and most of the information mentioned in these reports refers to something the protagonists said or wrote. By using direct speech and avoiding using personal interference from the author, the information is presented as factual and the author becomes an omniscient narrator: the author accurately knows what the people in the accounts did, said and even felt (10). (10)

Al punto sintió que le dieron una buelta los ojos y quedó con vista clara, viendo y conociendo distintamente a todos los presents. (Mun) [At that moment she felt that her eyes rolled, and her sight became clear, being able to see and recognise every person there]

In this example, the author explains what the protagonist did feel when the miracle took place, that is, when she recovered her sight. Interestingly, apart from the use of sight-​related verbs in the narrative, the miracle is also related to sight.

332

Natalia Mora-López

In opposition to direct speech, indirect speech refers to those cases where the author relates other person’s words without directly quoting them. For example, (11) illustrates one of these cases. (11)

Y cotejada con otras imágenes del mismo san Ignacio –​según ha referido un sacerdote que de allá a venido a esta corte–​tiene no sé qué cosa superior y extraordinaria. (Mun) [And checked against other statues also of saint Ignacio –​as a priest who has come to this court from that place has related–​ has I-​don’t-​k now-​what superior and extraordinary thing in it]

The reference does not seem very accurate: the reader is not given any information about this priest and the words reported are far from precise. The author may want to maintain the sense of unexplainability of the miracle or the miraculous qualities of the statue. 5.3. Indirect Evidence: Folklore Another source of information is folklore, although in this case, it is not a source from which the author of the account gets the information, but a source used by the protagonist of the narrative the author is presenting. In examples (12) and (13), the author introduces folklore as source for the witnesses’ initial knowledge of the miracles. (12) Sabiendo por fama los milagros que nuestro Señor obrava por medio de san Ignacio en Munébrega… (Mun) [knowing due to their popularity the miracles that our Lord worked through st. Ignacio in Munébrega…] (13) Con la fama de los milagros, determinó de visitar la capilla de san Ignacio (Mun). [Due to the popularity of the miracles, she resolved to visit the saint Ignacio’s chapel]

The author says here that the miracles were so popular that the protagonists knew about them and had an influence on their decisions.

Evidential Expressions in Spanish Miracle Accounts

333

Therefore, it can be inferred that it was part of the inhabitants’ shared knowledge and folklore. 5.4. Direct Evidence: Attested by Author Some examples of evidential expressions in the author’s voice are also found. The examples below illustrate the situations in which the events were perceived and transmitted by the author himself. (14)

Diéronselos a fuerça de sus peticiones y mejoró con ellos visiblemente. (Moy) [she was given them thanks to the strength in her petitions and they made her get better visibly]

It can be understood in (14) that visiblemente [visibly] refers to what the people surrounding the nun thought or saw, but it also emphasizes that the fact that she recovered from her illness, that is, the miracle, is something that everyone could see. Once again, the author insists on the repetition of terms related to the senses of hearing and seeing when reporting the miracles. In (15), the author uses the verb mostrar [show] to talk about the protagonist’s behavior. (15)

echaron mil bendiciones por la honra y labor que hizo a aquel convento y afecto que ha mostrado en premiar las singulares virtudes de nuestro padre. (Sim) [they blessed the honour and work that he did in that convent and the affection he has shown to reward the unique virtues of our father]

Therefore, the author presents his affection feelings as something that had been externalized and could, therefore, be perceived by the senses.

334

Natalia Mora-López

5.5. The Use of Sight and Hearing Together with reported verbs of saying, there is a prominent use of verbs related to sight and hearing, typically ver [see] and oír [hear]. Thus, most of the information mentioned in these reports refers to something the citizens said, saw or heard, as seen in (16) and (17). (16)

Un niño que estava con garrotillo, ya en el último trance de la muerte, con voz debilitada dixo: “Agüela, he visto subir al cielo al padre Roxas cercado de muchos ángeles y luzes”. (Sim) [A boy who suffered from dipthteria, in the last moments before passing away, said with a weak voice: “Granma, I’ve seen father Roxas ascent to heaven, surrounded by a lot of angels and lights”.]

(17)

Respondió el santo: “Ya la oigo, yo cumpliré lo que la he prometido”. (Sim) [Replied the saint: “I do hear you, I will do what I promised you”.]

In the case of direct speech in (16), the child says that he has seen something (the saint ascending to heaven). Therefore, the author is transmitting the experience of a first-​hand witness who has perceived the facts with his own eyes. In (17), it can be seen that it is not only the words of witnesses that are quoted but also the words of sacred entities, like saints. In fact, another reference to a saint illustrates the emphasis on sight, when a protagonist talks about a saint’s intentions regarding a miracle (18). (18) mi santo Francisco Xavier, que hasta en esto me quiso mostrar su grandeza pues me sanó dexándome dentro de mí la causa de mi mal que poco a poco fue saliendo en los sudores, para que yo lo viesse con mis ojos. (Moy) [my saint Francisco Xavier, who even in this wanted to show me his greatness since he healed me and left inside me the cause of my condition which would little by little sweat out of me, so that I would see it with my own eyes]

The protagonist says that the way the saint used to show he had worked a miracle was to make the devotee sweat his condition out and, consequently, see it.

Evidential Expressions in Spanish Miracle Accounts

335

In example (19), the author indicates that the protagonist heard someone relating the miracles performed by a saint, which made him decide to go and visit the saint’s chapel to show his devotion and pray for a miracle for himself. (19) Oyendo contar los milagros que san Ignacio obrava en Munébrega, fue a visitar su capilla… (Mun) [Hearing someone telling the miracles worked by Ignacio in Munébrega, he went to visit his chapel…]

Similarly, when direct speech is not used to let the protagonists say what they heard or saw by themselves, the author may indicate this inside the narrative, as in (20), (21), (22) and (23). In these cases, the author indicates that as a result of seeing or hearing of the miracles worked by a saint, their devotion increased or they decided to go and ask the saint for a miracle for them, with a positive outcome. (20) Viendo los milagros que san Ignacio hazía, con desseo de sanar cobró mucha devoción al santo. (Mun) [Seeing the miracles that saint Ignacio did, willing to recover health he became very devoted to this saint] (21) Inés Cavallero, muger de Martin de Huerta, vezinos de Nuévalos, estando ciega de suerte que no podía dar un passo sin guía (…), oyendo los milagros que san Ignacio hazía, se determinò de ir a Munebrega. (Mun) [Inés Cavallero, woman from Martin de Huerta, neighours of Nuévalos, being as blind as a guide was needed to take any steps (…), hearing of the miracles that saint Ignacio did, made up her mind to go to Munebrega] (22) huyó toda enfermedad sin dexar señal alguna de lo que avía sido, como lo vio y experimentó todo el convento y después lo declaró ante el illustríssimo y reverendíssimo señor don Miguel de Portugal. (Moy) [All the illness ran away without leaving a single vestige of what it had been, as the convent in full had seen it and experienced it and later declared it before the very illustrious and revered gentleman Mr Miguel from Portugal] (23) Destos sudores fueron testigos muchas personas eclesiásticas que lo vieron y uno dellos que fue el padre maestro fray Pedro del Villar… (Mun) [This sweating was witnessed by many clergymen who saw it and one of them who was father Pedro del Villar…]

336

Natalia Mora-López

In example (22), the author emphasizes that all the members in the convent had seen and experienced and, consequently, witnessed, the miracle and so they told the bishop. In (23), the author also mentions that the miraculous events were witnessed by many people and he even names one of those witnesses. 5.6. Expressions Related to Reliability Some expressions aim to show that the texts and the facts related in them are totally trustworthy. These expressions usually deal with meanings of certifying, approving and authenticity, and they may partially overlap with epistemicity, since the purpose of providing this information is to show that it is true. In (24), the author introduces a transcription of the letters written by the witness of the miracle and qualifies it as being fielmente [faithfully] presented. (24) pondremos aquí fielmente la copia de dos cartas (Moy) [we will faithfully present here the copy of the two letters]

Therefore, readers are told in advance that everything they read from that moment on, will be the first-​hand account of events and miracles witnessed and lived by the author of the letter exactly as they were related. In (25), the author refers to the doctors who proved that the condition of the experiencer of the miracle was fatal and nothing could be done. This is used to emphasize the difficulty, unlikeliness and even impossibility of the experiencer’s recovery, thus turning it into a miracle. (25) Finalmente, (…) afirmaron los médicos con certificaciones auténticas no poder aquella vida durar con tal complicación de males. (Moy) [Finally, (…) the doctors claimed with authentic certificates that that life could not last in such a severe condition]

Evidential Expressions in Spanish Miracle Accounts

337

During the narrative, the author may try to prove the validity of the information by referring to an external approval, as in (26). (26) Vistas las informaciones y aprobadas por tantos teólogos que con tanta rectitud las han mirado… (Sim) [Seen the information and approved by so many theologists that so uprightly have looked at them… ]

In this case, the author refers to qualified experts in the field and a rigorous verification process followed for the approval of the information. This increases trusthworthiness and also the importance of the events, since they deserved those experts’ attention. Similarly, information in the accounts as a whole may pass a previous stage of approval and verification, as it is told at the beginning or end of the texts (27–​28). (27) Autenticado y aprobado en la forma que lo manda el concilio tridentino por el illustríssimo señor don Miguel de Portugal obispo de Lamego del consejo de su magestad. (Moy) [Authenticated and approved as required by the Council of Trent by the very illustrious gentleman Mr Miguel of Portugal bishop of Lamego from the council of the king] (28) Relación de la colocación que se hizo al reverendíssimo padre maestro fray Simón de Roxas a dos días del mes de febrero deste presente año de 1625. Refiérense en ella algunos milagros de los que se han aprobado en sus informaciones. (Sim) [Account of the placement done to the very revered father Simón de Roxas on February 2nd, 1625. Related in them some miracles whose information has been approved]

These examples tell the reader they are guaranteed the information the author shares is absolutely true since it has passed the required validity standards. However, in (29) a different situation can be seen.

338

Natalia Mora-López

(29) Ha se recogido esta relación de una averiguación que recibieron cinco notarios reales y uno apostólico ante los iurados del lugar de Munébrega y de las cartas que el padre maestro fray Pedro del Villar del orden de san Augustín, el padre Gerónimo Alegre, retor del Colegio de la Compañía de Iesús de Calatayud y el padre Valerio Piquer y el padre Francisco Pérez de Culla, de la misma Compañía, an escrito a los padres de Madrid. (Mun) [This account has comprised an enquiry received by five royal and one Roman Catholic notaries before the juries of Munébrega and the letters that father Pedro del Villar of the Augustinian order, father Gerónimo Alegre, teacher of rhetorics at the School of the Society of Jesus of Calatayud, and father Valerio Piquer and father Francisco Pérez de Culla, of the same Society, have written to fathers from Madrid]

In this case, readers are told the specific source where the information comes from. It is two-​fold: on the one hand, part of the information was initially received and, consequently, accepted as true by notaries and juries (otherwise, it would not have gone any further); on the other hand, the rest of the information comes from letters written by members of the church themselves.

6. Distribution of Occurrences The types of occurrences discussed in the previous sections are not equally present or distributed in every text. Table 2 shows the quantita­ tive results for each of the categories in each of the three texts. The raw number of occurrences is shown first. Below each of the raw numbers is the percentage they stand for with respect to the total number of ocurrences in that specific text. In the last column, the total number of ocurrences is given per category as well as the corresponding percentage with respect to the total number of ocurrences in the three texts. The last row shows the representation of the total number of ocurrences in each of the texts with respect to the total number of ocurrences:

339

Evidential Expressions in Spanish Miracle Accounts Table 2.  Occurrence distribution per category in each text Sim

Moy

Mun

Total

1834 w

3429 w 3539 w

8802 w

Indirect evidence: Inferential-Author

1

2

0

3

%

3.57

4.55

-

2.86

Indirect evidence: Inferential-Witness

0

3

0

3

%

-

6.82

-

2.86

Indirect evidence: Reported speech-Quote

10

4

12

26

%

35.71

9.09

36.36

24.76

Indirect evidence: Reported speechRephrase

2

13

3

18

%

7.14

29.55

9.09

17.14

Indirect evidence: Folklore

0

0

3

3

%

-

-

9.09

2.86

Direct evidence: Attested by author

2

3

0

5

%

7.14

6.82

-

4.76

Sight and hearing

9

15

15

39

%

32.14

34.09

45.45

37.14

Reliability expressions

4

4

0

8

%

14.29

9.09

-

7.62

Total

28

44

33

105

%

26.67

41.90

31.43

100

It must be noted that one occurrence classified as Sight and hearing was also classified as Direct evidence: Attested by the author, due to its overlapping nature. This case is illustrated in (14) and refers to the use of visiblemente [visibly]. Although the number of words in Moy and Mun was similar (3429 and 3539, respectively), the number of occurrences differed to some extent (44 vs. 33). However, the shortest text (Sim, with 1834 words) did include fewer occurrences (28). Regarding the four major categories –​indirect evidence, direct evidence, sight and hearing, and reliability –​, there is a preference for the use of the first one, with about half the total number of ocurrences of the analyzed items in each of

340

Natalia Mora-López

the texts (46.42%, 50% and 54.55%, in Sim, Moy and Mun, respectively). The second most common resource is the use of sight and hearing expressions (32.14%, 34.09% and 45.45%). Neither Direct evidence nor reliability expressions were found in Mun, but the number of ocurrences for these two categories in Sim and Moy differed slightly. Indirect evidence has been seen to be the most used category. However, there are some specific subtypes that obtained the highest results in comparison to others. More specifically, it was reported speech that comprised most of the ocurrences of indirect evidence. The use of direct quotes or rephasing varies depending on the text: while Sim and Mun have a higher use of quotes (37.71% and 36.36%) and fewer rephrasing cases (7.14% and 9.09%), thus priotizing the protagonists’ own words, Moy shows opposite results (9.09% for quotes and 29.55% for rephrasing). Inferential evidence, either narrated by the author of the text or by a character in it (a witness), and Folklore evidence were scarcely present in the texts. However, while any type of inferential evidence appears in Sim and Moy, but not in Mun, it was only in Mun where Folklore evidence had any occurrences. With respect to Direct evidence, few cases were found, this being the least used resource and only present in two of the texts (Sim and Moy). This lack of occurrences may be related to the authors’ intention of presenting information as a fact and their role as reporters of events actually experienced by someone else. It could have been the case that some other characters in the accounts had transmitted their experiences through Direct evidence resources when their words or letters are transcribed, but this was not found either. It is not always the case that they are quoted and, when they are, they are more likely to use Indirect evidence resources. Expressions that may convey a viewpoint are generally avoided, but many instances of what people see and hear are introduced, as can be seen in the high frequency of these elements (32.14%, 34.09% and 45.45% in Sim, Moy and Mun). As shown in examples (16)-​(23), these verbs usually have noun phrases as complements to indicate what was seen and heard. It can be said that, together with indirect evidence, authors turn to this resource as a major tool to make readers feel that the events narrated did happen and were real.

Evidential Expressions in Spanish Miracle Accounts

341

Finally, expressions which were taken as evidence of the reliability of the accounts and miracles, typically related to prestigious entities certifying and approving the information, are also discreetly included in two out of the three texts (14.29% and 9.09% in Sim and Moy). However, since some of them are used to validate the information in the full document, their occurrence results may look smaller than the effect they produce.

7. Conclusions This study analyzed the use of evidential meanings in three accounts of religious miracles that were written in the 17th century. These accounts can be compared to modern newspapers or news and related events of common interest. The analysis divided the results of the types of evidential expressions into several categories: three types of indirect evidence (inferential, reported and folklore), and one type of direct evidence (attested by the author). Additionally, expressions related to senses, more specifically, sight and hearing, were analyzed and discussed as well, due to the important role they play in relation to evidential expressions. Finally, the analysis also covered the use of some strategies to increase the level of reliability of the information related, due to their relevance in the accounts and influence on evidentiality. It can be concluded that the lack of importance the author has at the time the accounts were written may be related to a scarce use of direct evidential strategies. Together with the fact of avoiding viewpoints not only from the author’s side but also from the characters’, this provides the narrative with a more objective style: things just happened as they (either authors or characters) say or write. For this reason, the use of indirect evidence, especially reported evidence, is the most common resource and the appearance of verbs of saying is profuse, since the witnesses’ words are rephrased or introduced explicitly in the texts. Some instances of indirect inferential evidence were found in the voice of either the authors or the witnesses. Additionally, indirect evidence

342

Natalia Mora-López

that makes use of folklore was also present in the texts, which was used to highlight the fact that miracles were known by most of the people in the place. When evidential expressions appear to report direct evidence, they do in the author’s words. However, these are not a frequently used resource, since authors do not usually experience the miracles by themselves and both authors and witnesses make use of indirect evidence more commonly. Expressions related to hearing and seeing are abundant in the three texts. Even though they are included in the narrative and usually are limited to indicating what someone specifically heard or saw, they are the grounds of the miracles. Indirect evidence and this type of expressions are the essential resources that authors use to emphasize that the events did actually happen. Finally, some less frequent strategies are also used to convince the reader of the truthfulness of the accounts. The ones analyzed here include specific lexical items ( fielmente [faithfully]), explicit and external verification and approval processes and the reference to members of the church as originally reporting sources of the facts. This set may overlap to some extent with epistemicity, since the purpose of these strategies is to prove that the information is true, that is, the miracles did actually take place. Despite these conclusions, it must be acknowledged that only three texts were studied. Although the analysis can be extended to the other six accounts in the larger corpus, the singularity of this type of texts makes them difficult to study quantitatively. However, considering the content of the accounts, other strategies and meanings may throw interesting results, like the use of epistemicity.

References Ahern, Aoife K. /​Amenós-​Pons, José /​Guijarro-​Fuentes, Pedro 2015. Metarepresentation and evidentiality in Spanish tense and mood. Belgian Journal of Linguistics 29/​1, 61–​82.

Evidential Expressions in Spanish Miracle Accounts

343

Aikhenvald, Alexandra Y. 2004. Evidentiality. Oxford: Oxford University Press. Aikhenvald, Alexandra Y. /​Dixon, R. M. W. (eds). 2003. Studies in Evidentiality. Amsterdam/​Philadelphia: John Benjamins. Bermúdez Wachtmeister, Fernando 2004. La categoría evidencial del castellano: metonimia y elevación de sujeto. Boletín de Lingüística 22, 3–​31. Bybee, Joan 1985. Morphology. A Study of the Relation between Meaning and Form. Amsterdam/​Philadelphia: John Benjamins. Carretero, Marta /​Zamorano-​Mansilla, Juan Rafael 2013. Annotating English adverbials for the categories of epistemic modality and evidentiality. In Marín-​Arrese, Juana I. /​Carretero, Marta /​ Arús Hita, Jorge /​van der Auwera, Johan (eds) English Modality: Core, Periphery and Evidentiality. Berlin: Mouton de Gruyter, 317–​355. Carretero, Marta /​Zamorano-​Mansilla, Juan Rafael 2015. Evidentiality as conversational implicature: Implications for corpus annotation. Procedia -​Social and Behavioral Sciences 212, 146–​150. Chafe, Wallace L. 1986. Evidentiality in English conversation and academic writing. In Chafe, Wallace L. /​Nichols, Johanna (eds) Evidentiality: The Linguistic Coding of Epistemology. Norwood NJ: Ablex, 261–​272. Chafe, Wallace /​Nichols, Johanna (eds). 1986. Evidentiality: The Linguistic Coding of Epistemology. Norwood, NJ: Ablex. Comrie, Bernard 2000. Evidentials: Semantics and history. In Johanson, Lars /​Utas, Bo (eds) Evidentials. Turkic, Iranian and Neighbouring Languages. Berlin/​New York: Mouton de Gruyter, 1–​12. Cornillie, Bert 2007. Epistemic Modality and Evidentiality in Spanish (Semi-​)Auxiliaries. A cognitive-​functional Approach. Berlin/​ New York: Mouton de Gruyter. Cornillie, Bert 2008. On the grammaticalization and (inter)subjectivity of Spanish evidential (semi-​)auxiliaries. In Seoane, Elena /​ López Couso, María José (eds) Theoretical and Empirical Issues in Grammaticalization. Amsterdam/​Philadelphia: John Benjamins, 55–​76.

344

Natalia Mora-López

Cornillie, Bert 2009. Evidentiality and epistemic modality: on the close relationship of two different categories. Functions of Language 16/​1, 44–​62. Cornillie, Bert 2015. Más allá de la epistemicidad. Las funciones discursivas de los adverbios epistémicos y evidenciales en el español conversacional. Spanish in Context 12/​1, 126–​144. Cornillie, Bert /​Gras Manzano, Pedro 2015. On the interactional dimension of evidentials. The case of the Spanish evidential discourse markers. Discourse Studies 17/​2, 141–​161. Dendale, Patrick /​Tasmowski, Liliane 2001. Introduction: Evidentiality and related notions. Journal of Pragmatics 33/​3, 339–​348. Ferrari, Laura /​Gallardo, Susana 1999. Los marcadores de evidencialidad empleados por la prensa en una controversia ambiental. Revista Iberoamericana de Discurso y Sociedad 1/​4, 69–​93. García de Enterría, María Cruz 1994. La Hagiografía popular barroca: entre lo maravilloso y lo fantástico, Draco. Revista de Literatura Española 3/​4, 191–​204. García de Enterría, María Cruz 1996. Relaciones de sucesos en pliegos de villancicos del siglo XVII. In Ettinghausen, Henry /​García de Enterría, María Cruz /​Infantes Víctor /​Redondo, Augustin (eds) Las relaciones de sucesos en España (1500–​1750). Actas del Primer Coloquio Internacional (Alcalá de Henares, 8, 9 y 10 de junio de 1995). Madrid: Publications de la Sorbonne and Servicio de Publicaciones de la Universidad de Alcalá, 167–​176. García de Enterría, María Cruz 1998. Magos y santos en la literatura popular (superstición y devoción en el Siglo de las Luces). In Huerta Calvo, Javier /​Palacios Fernández, Emilio (coords) Al margen de la Ilustración. Cultura popular, arte y literatura en la España del siglo xviii. Amsterdam: Rodopi, 53–​76. Givon, Talmy 1982. Evidentiality and Epistemic Space. Studies in Language 6/​1, 23–​49. González Ruiz, Ramón /​Izquierdo Alegría, Dámaso /​Loureda Lamas, Óscar. 2016. Un acercamiento a los fundamentos de la evidencialidad y a su recepción y tratamiento en la lingüística hispánica. In González Ruiz, Ramón /​Izquierdo Alegría Dámaso /​Loureda Lamas, Óscar (eds) La evidencialidad en español: Teoría y

Evidential Expressions in Spanish Miracle Accounts

345

descripción. Madrid/​Frankfurt am Main: Iberoamericana/​Vervuert, 9–​45. de Haan, Ferdinand 1999. Evidentiality and Epistemic Modality: Setting Boundaries. Southwest Journal of Linguistics 18, 83–​101. Laca, Brenda 2000. Matizaciones, modalizaciones, comentarios. In Vázquez, Graciela (coord.) Guía didáctica del discurso académico escrito. Madrid: Edinumen, 95–​105. López Ferrero, Carmen 2005. Funciones retóricas en la comunicación académica: formas léxicas de modalidad y evidencialidad. Signo y Seña 14, 115–​139. Marín Arrese, Juana I. /​Carretero, Marta 2014. The Expression of Evidentiality and Epistemic Modality in English and other European: A cross-​linguistic study of must and deber (de). In de la Cruz Cabanillas, Isabel /​Tejedor Martínez, Cristina (eds) Linguistic Insights: Studies on Languages. Alcalá de Henares: Universidad de Alcalá de Henares, 336–​346. Marín Arrese, Juana I. /​Hidalgo, Laura /​Molina, Silvia 2002. Evidentiality and Modality in English and Spanish: Connections and Interactions. In Facchinetti, Roberta (ed.) English Modality in Perspective. Genre Analysis and Contrastive Studies. Verona: Centro Arti Grafiche Universitá di Verona, 135–​146. Marín Arrese, Juana I /​Hidalgo, Laura /​Molina, Silvia 2004. Evidential, epistemic and deontic modality in English and Spanish: The expression of writer stance in newspaper discourse. In Facchinetti, Roberta and Palmer, Frank (eds) English Modality in Perspective. Genre Analysis and Contrastive Studies. Frankfurt am Main: Peter Lang Verlag, 121–​139. McCready, Eric /​Ogata, Norry 2007. Evidentiality, Modality and Probability. Linguistics and Philosophy 30/​2, 147–​206. Nuyts, Jan 2001. Epistemic Modality, Language, and Conceptualization: A Cognitive-​ pragmatic perspective. Amsterdam: John Benjamins. Palmer, Frank R. 2001. Mood and Modality. 2nd ed. Cambridge: Cambridge University Press. Sánchez Iglesias, Alicia 2017. Las relaciones de sucesos en la colección de Jesuitas de la Biblioteca de la Real Academia de la Historia (1600–​1650): catálogo completo junto con el estudio y edición

346

Natalia Mora-López

de las relaciones sobre milagros. Unpublished PhD dissertation. Salamanca: University of Salamanca. Vatrican, Axelle 2015. Evidentiality and epistemic modality in the rumor/​journalistic conditional in Spanish. Belgian Journal of Linguistics 29/​1, 83–​100. Willett, Thomas 1988. A cross-​linguistic survey of the grammaticalization of evidentiality. Studies in Language 12, 51–​97.

Liisa Vilkki

The Use of the English Modal Verbs in Linguistic and Philosophical Research Articles

Abstract This paper discusses how the English epistemic (could, may, and might) and inferential (must and should) modal verbs are used in linguistic and philosophical research articles published in internet journals. My theoretical framework is the metadiscourse framework, and I also draw on other studies in academic discourse and politeness. The main questions addressed are the following: What kind of uses of the selected modals can be distinguished in linguistic and philosophical research articles? Are there differences in the usage of the modals in linguistic and philosophical research articles? How can the possible differences be explained? Two electronic corpora of research articles were compiled to answer these questions. They were taken from five linguistic and five philosophical internet journals. The results of the analysis show that these modals often have similar, multifunctional uses in both types of articles. However, important differences were also found. Philosophical articles, especially, include a higher amount of various predominantly reader-​oriented usages of these modal verbs. The nature of philosophical research, which mainly involves re-​assessing prior positions and presumptions, can explain this difference. Philosophers must take special care to control the level of personality to build convincing arguments. Keywords: epistemic, inferential, hedge, booster, linguistic, philosophical, research articles

348

Liisa Vilkki

1. Introduction This paper’s purpose is to investigate how the five English epistemic (could, may, might) and inferential (must, should) modal verbs are used in linguistic and philosophical research articles, published in internet journals. The study focuses on both quantitative and qualitative disciplinary differences in the use of these markers. Academic research writing has traditionally been perceived to be focused on conveying factual information, and academic texts have been regarded as channels of communicating independently existing truths. Discourse analysts over the last three decades have argued in reaction to this traditional characterization that academic writing and academic texts embody interactions between writers and readers (e.g. Myers 1989; Hunston 1994; Hyland 1998a,b, 2005a,b; Charles 2006). Several researchers (e.g. Hun­­­­ ston 1994: 192–​193; Hyland 2005b: 173; Charles 2006: 492) argue that academic discourse seeks to persuade the reader to accept the author’s viewpoint or research findings, and thus should not be regarded as objective, faceless and impersonal. Knowledge and knowing are social institutions in this view, collectively created through the interactions of individuals. From this perspective, epistemic (e.g. could, may, might) and inferential (e.g. must, should) modal verbs and adverbs (e.g. perhaps, possibly, certainly, obviously) have a significant role in academic discourse. For example, Hyland (2005b: 173–​80) argues that these and other kinds of expressions are used to control the level of personality, and that they are central to building a convincing argument in academic texts. Epistemic expressions indicate some degree of the speaker’s certainty, while inferential expressions code at least some type of the speaker’s inference as an information source, as a kind of evidentiality (Vilkki, forthcoming). Inferential expressions may also combine both various inferential properties and degrees of the speaker’s certainty (e.g. must, should). It is important to distinguish these semantic definitions of the epistemic and inferential devices from their use as a rhetorical strategy in academic discourse. These devices allow writers to withhold complete commitment to a proposition or to express their certainty of a proposition, while they simultaneously help writers to

The Use of English Modal Verbs in Research Articles

349

convey deference for readers’ views, or to mark solidarity with readers. These kinds of multifunctional expressions are often called hedges (e.g. may, might) and boosters (e.g. must) (cf. Hyland 1998a,b, 2005b) in the studies of academic discourse. Hedges have especially been widely studied in different branches of pragmatics and applied linguistics during the last few decades (cf. articles in Kaltenböck/​Mihatsch/​ Schneider 2010b). General agreement exists today that hedging is a rhetorical or discourse strategy that basically “reduces the force or truth of an utterance and thus reduces the risk a speaker runs when uttering a strong or firm assertion or other speech act”. (Kaltenböck/​Mihatsch/​ Schneider 2010a: 1). Thus, the fact that an expression may be used as a hedge is not part of its definition (cf. Fraser 2010: 23). Hedging devices are drawn from every syntactic category, and we might say that hedges form an open functional class. Similarly, the possibility for an expression to be used as a booster is not part of its definition. Boosters allow writers to express their certainty in what they say and to indicate involvement with the topic and solidarity with their audience (cf. Hyland 2005b: 179). Hyland’s (1998a,b) studies of hedges and boosters represent the early studies of these expressions in academic discourse. Hyland (1998b) identified the major hedging devices used in cell and molecular biology research articles, and the overall findings are compared with other forms of academic writing. Hyland (1998a) investigates both hedges and boosters, and the data of this study comprise a corpus of research articles in eight disciplines together with interviews with members of the relevant discourse communities. For example, articles of philosophy and applied linguistics are included in the corpus. The quantitative results demonstrate the importance of both hedging and boosting in academic texts (Hyland 1998a: 356–​7). They also show that hedges exceeded boosters by nearly 3 to 1. Hyland assumes that the latter result reflects the importance of both distinguishing fact from opinion in academic discourse and the need for arguments to be presented provisionally rather than assertively. Hyland also found substantial differences in the disciplinary results. The results overall suggest a general division between philosophy, marketing, linguistics and sociology on the one hand, and physics and engineering on the other. Biology occupies the middle ground. This distinction is illustrated by the fact

350

Liisa Vilkki

that over 70 percent of all hedges occurred in the humanities/​social science articles, and they were over twice as frequent in philosophy, marketing and linguistics as in physics and engineering. However, there was more of a disciplinary imbalance in the number of boosters. Over a quarter of the total boosters occurred in philosophy and less than 7 percent in engineering. Hyland, however, concludes that the results show major disciplinary differences that reveal themselves as a clear distinction between the sciences and humanities/​social sciences. Hyland’s later studies (e.g. 2005b: 173–​80, 2005a: 49, 2015b: 1–​3) interpret the notions of hedge and booster as representing broader notions of stance and metadiscourse. Myers (1989), Barton (1993), Salager-​Meyer (1994), and Crompton (1997) are also among the early studies of hedging in academic discourse. Takimoto’s (2015) corpus study of hedges and boosters, representing eight disciplines, is inspired by Hyland’s (1998a) early study of these concepts. Among the disciplines are linguistics and philosophy. The results of this study corroborate the findings of Hyland (1998a). According to Takimoto (2015: 98), hedges exceeded boosters in all disciplines, and approximately 70% of all hedges and boosters occurred in the humanities and social sciences. Philosophy papers evidenced considerable uses of hedges and boosters. Hedges and boosters have been discussed in more recent studies of academic discourse in terms like stance, metadiscourse (cf. Hyland 2005a, b; 2015b above), evaluation, appraisal, and voice. Xie (2020) takes the concept of evaluation as a starting point and outlines the major strands of research on authorial evaluation in English academic writing. He also discusses the major methodological approaches in the field of applied linguistics over the past twenty years. His review identifies four broad strands in terms of the analytical framework applied: the stance strand, the metadiscourse strand, the appraisal strand, and the voice strand. Xie’s classification of literature in these strands shows that evaluation and the related concepts proliferate in studies of academic writing. Recent studies on authorial evaluation tend to cover two dimensions: the propositional and the attitudinal dimension (Xie 2020: 1–​2). The propositional dimension indicates speakers’/​w riters’ certainty of or commitment to the proposition in terms of reliability or trueness, and the attitudinal dimension indicates speakers’/​w riters’ positive or negative feelings toward entities.

The Use of English Modal Verbs in Research Articles

351

Hedges and boosters are, of course, linked to the propositional dimension. According to Xie (2020: 3–​5), the notion of stance in the stance framework is basically equivalent to the notion of evaluation, covering both the propositional and the attitudinal dimensions. For example, Chan (2015) applies the stance framework in his study of the expression of stance in dissertation acknowledgments. His corpus represents six disciplines, including applied linguistics, and his quantitative analysis shows that the soft disciplines use more stance features than the hard disciplines. Chan uses the notions of hedge and booster only when he refers to earlier studies. Hyland (2005a,b, 2015b) represents the metadiscourse framework. This framework dis­ tinguishes two dimensions of metadiscourse: the interactive dimension and the interactional dimension, with the latter being further divided into stance and engagement. Stance here corresponds to both the propositional and the attitudinal dimensions of evaluation. Self-​mention (e.g. the first-​person pronouns) also represents stance. Hyland’s (2005b: 186–​187) analysis of the research article corpus from eight disciplines shows that the articles of the more discursive soft fields of philosophy, marketing, sociology and applied linguistics contained the highest proportion of interactional markers with some 75 percent more items than the engineering and science articles. As was previously discussed, hedges and boosters are included in stance in Hyland’s (2005a,b, 2015b) metadiscourse framework. The results of Hyland’s (2005b) study are thus similar to the results of his earlier study (Hyland 1998a), also discussed before. The central difference between these two studies is that Hyland (2005b) uses a more com­ prehensive framework. Applying the Appraisal framework is the third strand of research (Xie 2020: 3–​4). This multilayered framework comprises subsystems and their subcategories. Within this framework, the attitude system corresponds to the attitudinal dimension of evaluation and the engagement system to the propositional dimension of evaluation. For example, Oliver’s (2015) corpus study represents the Appraisal framework. She focuses on research papers, case reports and book reviews and analyzes hedging expressions and attitude markers from a cross-​linguistic (English vs. Spanish) and cross-​disciplinary (medicine vs. linguistics)

352

Liisa Vilkki

approach. Oliver (2015: 144–​146) proposes a hedging taxonomy that differs from the usual interpretation of hedging in the studies of academic discourse. Four pragmatic categories are distinguished in this taxonomy: shields (e.g. modal verbs, probability adverbs), approximators (e.g. adverbs of quantity and degree), author’s personal doubt and involvement (conditional, first-​person markers), and agentless strategies (e.g. passive voice). Only the first two categories, shields and approximators, correspond to the usual interpretation of hedging in the studies of academic discourse. The voice strand, unlike the other three strands, does not have a unified analytical framework (Xie 2020: 4–​5). This might be due to the broad range of meanings ascribed to the notion of voice (cf. Tardy 2012: 34). According to Tardy (2012), voice encompasses three dimensions: individual, social, and dialogic. The individual aspect represents writers’ “unique and recognizable imprint”, and it is realized via linguistic features such as rhetoric and evaluation (Tardy 2012: 37). Thus, the individual aspect of voice overlaps with the attitudinal and propositional dimensions of evaluation. The social aspect of voice is related to writers’ disciplinary and social contexts, because their choices of discursive features are shaped by the possibilities provided by the social contexts they are situated in. Voice as dialogic involves the interaction between the individual and the social dimensions, especially between writers and readers. A writer’s unique position in the discourse and the dialogic involvement of other voices may be seen as constitutive of authorial identity (cf. Bondi 2012: 102). For example, Gross/​Chesley (2012) investigate hedging in medical research articles. They regard hedges (possible, may) as an aspect of voice and discuss the relation of voice to stance. Vold (2006: 228) also discusses the concept of voice in her study of epistemic markers in research articles from linguistics and medicine. Her study is also inspired by, for example, Hyland’s (1998a) earlier work. Vold (2006: 225) states that markers of epistemic modality consti­ tute a dominant and basic type of hedges. Her findings show that both linguistics and medicine use epistemic modality markers to cautiously present conclusions, to suggest possible explanations, to put forward hypotheses and to indicate the study’s limitations (Vold 2006: 244–​246). The latter function, however, is more frequent in

The Use of English Modal Verbs in Research Articles

353

medical articles. However, epistemic modality markers in linguistic articles are often used in contexts of overt argumentation and are also often used to mitigate criticism of other research. They are also used to express caution when interpreting other researchers’ work. This function was not found in the medical articles. Voice represents the interactional dimension consisting of stance and engagement in Hyland’s (2005a: 49–​50) metadiscourse framework. More exactly, stance and engagement may be thought of as strategies for expressing voice through the ways writers use community-​sensitive linguistic resources to represent themselves, their positions and their readers. Hyland’s conception of voice may be characterized as encompassing all the three dimensions: individual, social, and dialogic (see Section 2). The interpretations of the notions of hedge and booster usually seem to be similar in recent studies of academic discourse, representing different frameworks. These interpretations concur with the definitions previously presented in this section. However, other kinds of conceptions of hedge (and booster) have sometimes been presented, such as Oliver’s (2015: 144–​146) taxonomy, discussed earlier (cf. also Hinkel 2005: 29). The overview of previous studies of hedges and boosters in academic discourse suggests, for example, that a clear distinction exists between the sciences and humanities/​social sciences in the use of these devices. It has been shown that hedges and boosters are particularly strongly represented in philosophy papers. Previous studies, however, mainly discuss both philosophy and linguistics as representing the humanities and social sciences. The differences between these disciplines in the use of hedging and boosting strategies have not, to my knowledge, been investigated in detail. This paper aims to examine the use of the modal verbs could, may, might, must, and should in the research papers from these two disciplines to address this gap. Other markers that are used as hedges and boosters in these disciplines should be examined in future studies. The present paper poses the following research questions: What kind of uses of the selected modal verbs can be distinguished in linguistic and philosophical research articles? Are there differences in the usage of the modal verbs in linguistic and philosophical research

354

Liisa Vilkki

articles? How can the possible differences be explained? Section 2 discusses some important theoretical questions in this study. Section 3 presents the corpora compiled for this study. Section 4 deals with the results of the quantitative and qualitative analysis of the data. Finally, Section 5 focuses on the discussion of the results and their explanations, as well as some future prospects for the study of hedging and boosting in linguistic and philosophical research articles.

2. Theoretical Framework 2.1. Academic Discourse, Disciplines, Genres I assume that academic discourse, such as research articles, may be understood as having two interactional goals: epistemological and interpersonal/​relational (cf. Hyland 2004). Academic knowledge is socially created because the constraints on justifiable knowledge are socially constructed among members of the disciplines. Disciplines, such as linguistics and philosophy, can be interpreted as communities of practice. The notion ‘community of practice’ emphasizes situated activity and “a set of relations among persons, activity, and world, over time and in relation with other tangential and overlapping communities of practice” (Lave/​Wenger 1991: 98). The idea of disciplines as commu­ nities of practice implies a certain degree of interdisciplinary diversity and a degree of intradisciplinary homogeneity. According to Hyland (2015a: 35), the epistemological and social character of disciplines are both reflected and created through their rhetorical conventions; therefore, the distinctions between disciplines can be informed by the study of their rhetorical practices. Discipline-​ specific language choices help construct persuasive and valid arguments, and they also contribute to the reinforcement of group membership and disciplinary practices. Hyland (2015a: 33) makes a useful distinction between the notions of proximity and positioning, which is the fit between rhetorical conventions and the persona one wishes to project. Proximity concerns the relationship between the self

The Use of English Modal Verbs in Research Articles

355

and community, whereas positioning is about the relationship between the speaker and what is being said. When working in a discipline, the members need to proximate to the rhetorical conventions it routinely employs to claim membership and learn how to use these conventions to take positions on matters that the community values. The members can use the options available to position themselves in terms of personal stance and interpersonal alignments. Genres such as research articles play a key role from this perspective because they are community approved and personally meaningful discourses (cf. Hyland 2015a: 33). Genres can be understood as community resources that allow users to create and read texts with some assurance that they know what they are dealing with. Genres enable the particularities of any situation to connect with wider norms and practices. Genre constraints are, at the same time, the enabling conditions for originality. According to Hyland (2004: 12), the common practices in knowledge-​creating genres such as the research articles will involve at least the need to: • • • • •

establish the novelty of one’s position make a suitable level of claim acknowledge prior work and situate claims in a diciplinary context offer warrants for one’s views based on community-​ specific arguments and procedures demonstrate an appropriate disciplinary ethos and willingness to negotiate with peers

This list shows that the means by which authors present knowledge claims and account for their actions in research articles involve not only epistemological factors but also social and affective elements. The most productive way to study genres is through corpora, because they reveal the disciplinary conventions in the repeated patterns of language use (cf. Hyland 2015a: 34). Moreover, by showing how linguistic expressions are typically used in contexts, corpora offer evidence of authors’ orientations to disciplinary communities and the ways they stake out individual positions. Hyland (2015a: 34, 42) argues that corpus analyses of genres can reveal enactments of identity. Genres privilege certain ways of making meanings, so they encourage

356

Liisa Vilkki

the performance of certain kinds of professional identities. They also simultaneously provide the boundaries within which those identities are valued. According to Hyland (2015a: 42), the concepts of proximity and positioning are not synonymous with identity, but they may provide the most immediate access to its rhetorical construction because they focus on what individuals do to project themselves as credible academics within a shared professional context. Section 1 discussed previous studies of hedging and boosting strategies (e.g. Hyland 1998a, 2005b; Takimoto 2015) that suggest that a clear distinction can be made between the use of these strategies in the sciences and in the humanities/​social sciences. According to Hyland (2005b: 187–​89), the differences in rhetorical practices are inextricably related to the purposes of the disciplines. Natural scientists tend to see their goal as producing knowledge able to withstand the rigours of falsifiability and developed through relatively steady cumulative growth. Problems emerge in an established context so that readers are often familiar with prior research and that the novelty and significance of contributions can easily be recognized. The humanities and social sciences, in contrast, are more interpretative and less abstract. There are also greater possibilities for diverse outcomes and less control of variables, so writers must express their evaluations and work harder to establish an understanding with readers. The research papers examined in this study are from linguistics and philosophy, so it is important to briefly characterize the similarities and differences between these disciplines. The classification of disciplines into branches of science is, of course, always difficult (cf. Fløttum 2006: 20–​1; Hyland 2005b: 187–​88). For example, linguistics is often described as being situated at the crossroads of the humanities and the social sciences or even as extending across these branches and the natural sciences. The three branches represent knowledge bases with different characteristics. The natural sciences have a relatively homogenous and objective knowledge base, whereas the humanities are in many contexts considered to be hermeneutical, with subjective interpretation as a prominent feature. The social sciences occupy a middle position between the natural sciences and the humanities, with some of the knowledge base rooted in more objective natural science matters, while other parts represent more subjective, interpretive aspects. Of

The Use of English Modal Verbs in Research Articles

357

course, it cannot be said that knowledge claims made within linguistics are based solely on subjective considerations, because linguistics is certainly also characterized by regularities and some common methodological tools. However, the humanities researchers may, to a larger extent than other researchers, be said to create new knowledge through their arguing for relations between facts and notions. It is also difficult to classify philosophy within some specific branch of science. Previous studies of academic discourse (e.g. Hyland 1998a, 2005b; Takimoto 2015) have classified it as representing the humanities and the social sciences. Some philosophers, however, believe their task is similar to scientific inquiry, forming hypotheses, then testing and disseminating findings for wider discussion among peers (cf. Hobbs 2013: 29). Philosophers, in contrast to natural scien­ tists, spell out their arguments from personal reflection rather than empirical evidence (cf. Chudnoff 2007). They also use more discursive details, taking great care to create understanding with their readers. The central concepts in philosophical writing are clarity and precision. One of the features of these concepts may be the avoidance of language associated with personal interests, feelings, and commitments (cf. Hobbs 2013: 29). Both linguistics and philosophy may be character­ ized as discursive disciplines, but philosophy seems to be more discursive than linguistics. Philosophers must place emphasis on carefully dissecting research problems into their parts, only after which they can be resolved. A problem can be dissected into its parts and understood completely by writing with sufficient clarity. Another difference between linguistics and philosophy is that linguists may use empirical evidence in their studies, whereas philosophers usually present their arguments from personal reflection. 2.2. Epistemic Modality and Inferentiality Based on my large-​ scale typological study (Vilkki, forthcoming), I assume that various epistemic meanings form the semantic domain of epistemic modality that is concerned with the property ‘the speaker’s certainty’, which has several degrees. Thus, speakers indicate their degree of certainty of (the truth) of the proposition (e.g. the English

358

Liisa Vilkki

modal verbs could, may, might) with epistemic modality. Different kinds of evidential meanings occupy the semantic domain of evidentiality. Evidentials express an information source. Inferential meanings indicate the type of inference as information source, and they are included in the semantic domain of inferentiality. The property ‘the speaker’s degree of certainty’ may also characterize evidential meanings, such as the meanings of visual or reported evidentials. Inferential meanings are purely evidential insofar as they only indicate the speaker’s inference as a type of information source. Inferences are based on evidence of some kind, for example, general knowledge or experience, observable results, or sensory evidence. The type of evidence may also be left unspecified. Inferential meanings may also combine both evidential and epistemic properties. These types of inferential meanings cover the fuzzy boundary area between the domains of epistemic modality and evidentiality. These meanings usually emphasize more either the notion of inference or the notion of the speaker’s certainty. For example, the English modal verbs must and should express both inferential and epistemic properties. During the past two decades, different kinds of views of the relationship between epistemic modality and evidentiality have been proposed, especially in typological studies (e.g. Aikhenvald 2006; Boye 2012; Palmer 2001; Vilkki, forthcoming). Dendale and Tasmowski (2001) distinguish between three major viewpoints concerning the relationship between these semantic domains that can be found in modern studies. The authors call the first viewpoint ‘disjunction’: the domains are conceptually distinguished from each other. The second approach is ‘inclusion’: one of the domains is interpreted as falling within the semantic scope of the other. The third is ‘overlap’: the domains partly intersect. Dendale and Tasmowski clearly align themselves with the ‘overlap’ view. They state that diverging opinions concerning the relationship question are largely due to the empirical fact that in many languages, the evidential forms, coding the source of information, also indicate the speaker’s assessment of the information’s reliability. However, they do not argue that the overlap between the two domains is attested in every language. The position taken in this study is that different kinds of overlap are possible between evidentiality and epistemic modality. Regarding

The Use of English Modal Verbs in Research Articles

359

inferential meanings, they can be predominantly evidential or predominantly epistemic. However, pure inferential and pure epistemic meanings also exist. The modal verbs in many languages may also express other types of meanings. Coates (1983: 21–​30) makes a distinction between epis­ temic and root modal verbs. The latter types of modal verbs express meanings like ‘obligation’ (e.g. must, should), ‘permission’ (e.g. can, may), ‘possibility’ (can, may) and ‘ability’ (e.g. can). Other terms have also been used instead of root modality, sometimes with slight modifications in the definitions. For example, Palmer (2001: 70–​85) distin­ guishes event modality, which has the subtypes of deontic and dynamic modality, while Heine (1995) and Bybee/​Perkins/​Pagliuca (1994) prefer the term ‘agent-​oriented modality’. 2.3. Hedges and Boosters Section 1 discussed that it is possible to distinguish several strands of research on authorial evaluation in academic discourse. The framework this study follows is the metadiscourse framework proposed by Hyland (2005a,b, 2015b). This comprehensive framework, as briefly described in Section 1, regards metadiscourse as the ways writers refer to the text, the writer or the reader, and acknowledges the contextual specificity of metadiscourse (cf. Hyland 2005a: 48–​50). The interactive dimension in this framework represents resources that allow the writer to manage the information flow (cf. Tse/​Hyland 2006: 180–​181). It is concerned with ways of organizing discourse to establish the writer’s preferred interpretations and to anticipate the readers’ knowledge. The interactional dimension, expressing the writer’s textual voice, focuses on the participants in the interaction and reveals the extent to which the writer works to jointly construct the text with readers. Metadiscourse is thus an important means of facilitating communication, supporting a writer’s position and building a relationship with an audience. The exuberance of studies applying the metadiscourse framework has proved its vitality in revealing the ways that academic writers express evaluation (cf. Xie 2020: 3). Of course, this framework’s most relevant concepts for this study are hedges and boosters. They are included in stance,

360

Liisa Vilkki

one of the resources of the interactional dimension. Several epistemic and inferential modal verbs and adverbs (e.g. may, might, could, should, perhaps, possibly, probably, apparently, evidently) and other types of expressions (e.g. in my view, I think, seem(s), somewhat) have often been discussed as belonging to hedging strategies (e.g. Hyland 1998a,b, 2005a, b; Mauranen 1997, 2012). Hedges have been consid­­­ ered an important interactional feature and a rhetorical means in various academic genres in several disciplines, for example, in linguistics and philosophy. Hedges have been shown to frequently express several meanings simultaneously (e.g. Mauranen 1997; Hyland 1998a,b; Vold 2006; Diani 2014). Writers may use them to distinguish fact from opin­ ion, to offer a credible representation of themselves and their work, and to give polite deference to colleagues’ views. Moreover, several epistemic and inferential modal verbs and adverbs (e.g. must, certainly, of course, surely) and other types of expressions (e.g. indisputable, proves, undeniable) have been interpreted as boosters, which may also indicate several meanings simultaneously (e.g. Hyland 1998 a, b). Boosters allow writers to express conviction. They also mark an affective involvement and solidarity with an audience, stressing shared information, group membership, and direct engagement with readers. According to the corpus-​based studies of Hyland (1998a, b, 2005b), hedges and boosters index disciplinary practices, with both occurring more frequently in published papers in the humanities and the social sciences than in science and engineering fields. Hedges not only protect writers from unconsidered statements by implying that they are based on plausible reasoning rather than certain knowledge, but they also open discursive space for readers to dispute these statements. Writing in the more discursive fields involves far more recognizing alternative interpretations, because there is less control of variables, more diversity of research outcomes, and fewer clear bases for accepting claims than in the sciences and engineering. Writers do not suppose that readers will readily share their views, so they express their arguments more cautiously by using more hedges. However, because alternative interpretations are possible, writers also use more boosters in some circumstances to establish the significance of their work against other types of interpretations.

The Use of English Modal Verbs in Research Articles

361

The preceding discussion of hedges and boosters shows, for example, that they are often multifunctional in academic discourse. This discussion includes notions such as deference, solidarity, and involvement, which are often dealt with in politeness and face studies. I will briefly consider some relevant studies in these research fields in the next ­section. 2.4. Politeness and Face The notions of politeness and face have been among the most heavily debated notions in pragmatic and sociolinguistic research during the last decades (cf., for example, Watts 2003; Vilkki 2006; Bargiela-​ Chiappini/​Haugh 2009). It is beyond the limits of this paper to give an overview of politeness-​and face-​related research. Instead, I will briefly consider here Spencer-​ Oatey’s (e.g. 2005) rapport management (or politeness) approach, because I have found this approach useful in the present study. According to Spencer-​Oatey (2005: 9), rapport refers to the relative smoothness and harmony of relations between people, and rapport management is concerned with the management (or mismanagement) of relations between people. Spencer-​Oatey (2005) distin­­ guishes three key elements that influence people’s dynamic perceptions of rapport: behavioral expectations, face sensitivities, and interactional goals. Politeness is most clearly associated with behavioral expectations, and politeness is understood as the subjective judgments that people make about the social appropriateness of verbal and non-​verbal behavior. Spencer-​Oatey (2015) explains politeness in another publica­ tion essentially in terms of face and behavioral expectations, as well as the social conventions associated with these expectations. Face issues and behavioral expectations are different concepts, but they can also be connected. Face is concerned with notions such as status, reputation, competence, honor, dignity, and respect (cf. Spencer-​Oatey 2007: 651; O’Driscoll 2017: 102). If we feel that we are not valued in some way, we may perceive this treatment as face threatening. For example, if a person explicitly criticizes another person, the criticized person may feel her face is threatened. People are typically very face sensitive to

362

Liisa Vilkki

comments associated with their self-​aspects, for example, their competence, abilities, and group memberships (Spencer-​Oatey 2007: 643–​644, 2009: 141–​142). Spencer-​Oatey (2007) proposes that the effective study of face probably needs to take an identity perspective, for example, self-​aspect/​attribute approaches to the identity and values, associated with these attributes. Different kinds of face management strategies can be distinguished. It is possible to consider, in addition to the face-​ threatening strategy at least face-​protecting, face-​enhancing, and face-​ maintaining strategies (e.g. Wang/​Spencer-​Oatey 2015; O’Driscoll 2017: 97–​101). Another factor that can complement the explanation of politeness in terms of face sensitivity is the impact of behavioral expectations and the social conventions associated with them (cf. Spencer-​Oatey 2015). We may consider communities of practice, such as disciplines, from this perspective. We saw earlier in this section that disciplines develop conventions and frequently used patterns of language use that lead to expectations concerning the patterns that are appropriate in different contexts. If these expectations are not met, that can affect the judgments of the members of communities about politeness and influence the smoothness of their relations. The explanation of politeness in terms of behavioral expectations and social conventions is similar to Terkourafi’s (2015) association of politeness with conventionalization (cf. also Terkourafi/​Kádár 2017). However, Terkourafi’s (2015) focus is more on conventional expressions than on conventional behavior, broadly conceived. Terkourafi argues that conventionalized expressions can be adapted to a wide range of frequently experienced situations with minimal effort, and they are the most expedient means of achieving politeness or impoliteness. Departing from conventionalized expressions is, however, also possible, and may be associated with either increasing politeness or increasing impoliteness. Terkourafi presents different types of data, for example, findings of corpus studies, which converge on the idea that conventionalization is central to politeness. Conventionalization is inherently evaluative, “the right way of doing something” in some type of context.

The Use of English Modal Verbs in Research Articles

363

3. Data and Method The material of this study comprises 60 research articles. 30 articles were taken from five linguistic internet journals, and 30 articles were taken from five philosophical internet journals. All the journals were published either in 2015 or in 2016. The linguistic publications (277, 661 words) are Applied Linguistics (L_​A L), Discourse & Society (L_​ DS), Journal of Pragmatics (L_​J P), Language (L_​L), and Language & Communication (L_​LC). The philosophical publications (239, 317 words) are Ethic Theory & Moral Practice (P_​ETMP), European Journal of Philosophy (P_​EJP), International Journal of Applied Philosophy (P_IJAP), International Journal of Philosophical Studies (P_IJPS), and Philosophical Explorations (P_PE). The data for this study thus comprise two corpora –​ the linguistic corpus and the philosophical corpus. The two corpora were processed using a C+​+​program (Hart­ zell 2018) to find the frequencies and concordances of the selected modal verbs could, may, might, must, should. These modal verbs were selected because their frequencies were higher than the frequencies of the other English modal verbs. Of course, will and would show high frequencies, but they were not considered because they have a complex relationship to epistemic modality. More generally, the C+​+​program finds out the number of times an expression appears in the data, and it gives all the sentences in which this expression appears. Given the highly contextual nature of hedging and boosting, every instance was examined in its co-​text (often in the corpora, and not only in the concordances) to determine its actual use. Thus, this study draws on both a qualitative and quantitative approach to corpus analysis.

364

Liisa Vilkki

4. Results Section 1 posed the following research questions: What kind of uses of the selected modal verbs can be distinguished in linguistic and philosophical research articles? How can the possible differences be explained? I will first present the results of the quantitative analysis of the data in this Section, and then I will proceed to the qualitative analysis. The qualitative analysis is more important from the point of view of answering the questions. 4.1. The Quantitative Analysis Table 1 gives the frequencies for the selected five expressions (could, may, might, must, should). Given the imbalance between corpora sizes, relative frequencies are more important than absolute numbers. Occurrences per thousand words are considered clear enough in the present study. Table 1.  The frequency of the selected markers in the material. Expressions

Linguistic research articles: Occurrences f/1000 (no)

Philosophical research articles: Occurences f/1000 (no)

could may might must should

0.64 (177) 1.52 (422) 0.46 (128) 0.32 (90) 0.62 (171)

1.09 (261) 2.19 (523) 1.89 (451) 1.15 (276) 1.44 (345)

Total

3.59 (988)

7.76 (1856)

It is important to note that the numbers in Table 1 include all the uses of the selected modal verbs that were found in the corpora. This means that they also include root uses, hypothetical uses, and quasi-​subjunctive uses. All these uses are not considered in this study’s

The Use of English Modal Verbs in Research Articles

365

qualitative analysis, because the focus is placed on epistemic, inferential, and polite uses. Root modality is a cover term for different kinds of non-​epistemic modality, e.g., obligation, permission, and ability (cf. Section 2). Both epistemic and root hypothetical uses can be interpreted as representing conditionality (cf. Vilkki, forthcoming), and quasi-​subjunctive uses (e.g. ‘whatever this may/​might mean…’) also do not represent epistemic modality. Most of the occurrences of the modal verbs could, must, and should in the two corpora represent root modality. Could is also used to convey hypothetical root meanings. Only sometimes do they express epistemic or inferential meanings or merge these properties with root properties in their meanings. The philosophical articles contain a greater amount of root uses of all the modal verbs in comparison with the linguistic articles. This partly explains the higher number of these three modal verbs in the philosophical articles. The higher number of root uses in the philosophical articles also partly explains, why occurrences of may and might are more frequent in them than in the linguistic articles. The large-​scale corpus-​based study of the English modal verbs by Coates (1983) reveals that may and might are mainly epistemic, although they also express other types of meanings (Coates 1983: 131–​ 168). My preliminary qualitative analysis of the concordances of these modal verbs in the two corpora suggested that they are most often used as hedges. That is, they mostly seemed to express epistemic and polite properties. I counted all the different uses of may and might that I found in the corpora to clarify whether this assumption is correct. Tables 2 and 3 show the results: Table 2.  The frequency of the uses of may in the material. may

Linguistic research articles: Occurences f/1000 (no)

Philosophical research articles: Occurences f/1000 (no)

epistemic root concessive quasi-subjunctive

1.25 (347) 0.20 (56) 0.04 (11) 0.03 (8)

1.51 (363) 0.40 (95) 0.25 (59) 0.03 (6)

366

Liisa Vilkki

Table 3.  The frequency of the uses of might in the material. might

Linguistic research articles: Occurences f/1000 (no)

Philosophical research articles: Occurences f/1000 (no)

epistemic root (hypothetical) concessive epistemic hypothetical quasi-subjunctive

0.36 (100) 0.05 (15) 0.01 (4) 0.03 (9) 0 (0)

1.12 (287) 0.31 (73) 0.30 (72) 0.07 (16) 0.01 (3)

Epistemic and concessive represent hedging strategies in Tables 2 and 3. We saw in Section 2 that hedges are often multifunctional: They are used, for example, to indicate both cautious assumptions and polite deference to alternative views. Still, it is possible to claim that either the epistemic degree of certainty of assumptions (epistemic) or the polite property (concessive) may be predominant in the various hedging strategies. The epistemic and concessive uses are discussed and illustrated in the upcoming qualitative analysis. We can see in Tables 2 and 3 that both may and especially might have more epistemic and con­ cessive uses in the philosophical articles than in the linguistic articles. However, these uses together are more frequent than other types of uses in both corpora. The quantitative analysis shows first that all the selected modal verbs occur more frequently in the philosophical corpus than in the linguistic corpus. These frequencies included all the uses of the markers. Seven uses among the markers were distinguished: epistemic, inferential, root, quasi-​subjunctive, epistemic hypothetical, root hypothetical, and concessive. The philosophical corpus, in comparison with the linguistic corpus, contains a greater amount of root uses of all the modal verbs. This partly explains the higher number of these markers in the philosophical corpus. It was also found that most of the occurrences of could, must, and should in the two corpora represent root modality, whereas may and might are mostly used to express epistemic and polite (concessive) meanings, when these hedging strategies are counted together. Both may and especially might have more epistemic and concessive uses in the philosophical articles than in the linguistic articles.

The Use of English Modal Verbs in Research Articles

367

4.2. The Qualitative Analysis I present the results of the qualitative analysis of the material in this Section. I first discuss and illustrate some uses of the modal verbs could, must and should. However, the main focus is placed on the uses of may and might, which are most interesting from the perspective of the relationship between epistemic and polite properties. The modal verbs and other types of hedging expressions are marked in bold in the examples. The quantitative analysis mentioned that could, must and should are mostly used to express root (hypothetical) meanings in both corpora. The examples (1)–​(3) illustrate how these modal verbs may also be used as hedges (could, should) or boosters (must). Example (1) shows how the author uses could, first, to express an epistemic weak possibility. He claims that the concept ‘possible family’ should also involve weakly possible cases. He further specifies his claim by means of the adverb unlikely; this shows readers that he is really discussing only marginal possibilities. Could and (however) unlikely are used together as a hedging strategy. These markers are used by the author not only to convey epistemic weak possibility but also to avoid overstating his claim and risk inviting its rejection. The author thus presents his claim in a cautious manner. The caution or modesty protects the writer against the negative consequences that might arise if, at a later stage, the claims should turn out to be inaccurate. This hedging strategy also allows the author to open a discursive space in which readers can dispute his interpretations. The caution can be interpreted in terms of face. The writer protects his own face against the potential negative consequences, and he also protects the readers’ face, expressing a degree of deference to other possible opinions. Neither the epistemic nor the face-​protecting property is predominant in this example. (1)

Second, the contention that the concept ‘possible family’ is simply another way of grounding obligations arising from specific relations to human communities –​including the duty not to kill and eat a family member –​ overlooks the fact that it is not only dogs already belonging to families who fall under this richer narrative-​style concept of a fellow creature; it is any dog whatever who could become family, however unlikely that is. (P_​ETMP)

368

Liisa Vilkki

Examples (2) and (3) illustrate the use of must as a booster. It is used in both examples to assert the claim with confidence and to express a strong conclusion. Example (3) especially indicates involvement with readers, stressing shared information, because must is included in a question, directed to the audience. Shared information is implied by the author in example (2). He employs must to imply that, based on the preceding conditional sentence, the conclusion is necessary for anyone. Involvement with readers predominates in (3), whereas confident inference is foregrounded in (2). Both examples may be described as representing face enhancing. That is, the author enhances readers’ face in expressing shared information and group membership. (2)

If we value the presence of friends at our time of death –​when nothing is any longer in prospect for us –​then the values that we assign to friendship cannot be entirely prospective ones. Something must therefore be missing from any theory according to which the value of friendship is realized only in the future. This creates a difficulty for several theories, beginning with the work of Aristotle. (P_​IJPS)

(3)

Also, if … the presupposition of free will (FW) grounds (is sufficient reason for) faith in the existence of God and immortality (G&I), and free will is the necessary condition for the moral law (ML), then must not God and immortality be necessary conditions for the moral law, too? This is because if ‘ML entails FW’ (entailed by ‘not-​FW entails not-​ML’) and ‘FW entails G&I’ then ‘not-​G&I entails not-​ML’. (P_​ETMP)

The author uses should to convey probable conclusion in example (4) while simultaneously using it to show that he cannot infer for sure that readers will find his interpretation helpful. Thus, should simultaneously expresses not only a degree of certainty (probability) and inference but also modesty and caution. It is difficult to say whether the epistemic-​inferential property or the properties of modesty and caution could be interpreted as predominant. Neither of them is clearly on the foreground. Like example (1), this example may also be cast in terms of face. The author protects both his own face and his readers’ face. This hedging strategy also allows readers to reject the author’s conclusion.

The Use of English Modal Verbs in Research Articles

369

(4) Seeing this political-​philosophic point, in the way we have seen it, should help us think more generally about the ethics of medical communication. (P_​ETMP)

Example (5) represents a merger of epistemic-​ inferential (a probable conclusion) and root (attenuated obligation) uses. That is, it is used to indicate both the probable conclusion that “readers consider uncontroversial that …” and the view that “readers are obliged to consider uncontroversial that …” There is also some modesty and caution included in this use, but the merger properties seem to predominate. (5) How does an idealized version of P differ from the actual P? Well, it should be uncontroversial that idealized P differs in at least this respect: She has every true factual belief relevant to the decision at hand, and no false ones. Ideally, decisions are based on actual facts rather than false beliefs. Now, whether certain facts are or are not relevant is a normative question. (P_​ETMP)

Could, must, and should are mostly used in the corpora to indicate root meanings and only a few examples of epistemic, inferential, and polite uses were found; thus, it is not possible to describe the differences in the usage of these markers in linguistic and philosophical research articles. Future research on bigger corpora could reveal possible differences. The rest of the examples to be considered represent the most common types of uses of may and might found in the two corpora. The most frequent usage of these hedges in both corpora is to present claims in a cautious manner. This strategy was discussed earlier (see examples (1), (4) and (5)). Examples (6)–​(8) illustrate this usage. May and might also express possibility, quite a low degree of the speaker’s certainty, in these kinds of examples. Neither caution nor possibility clearly predominates, although a more detailed analysis could reveal fine contextual differences. (6)

The ‘default’ proximization strategy operating at the synchronic level is spatial proximization (often featuring a temporal component), whose function is to raise fears of imminence of the threat, which might be ‘external’ apparently but could materialize within the American borders virtually at anytime. (L_​DS)

370

Liisa Vilkki

(7)

This means that in this case there is consensus on the claim (a share in the fruits of social cooperation), but contestation on the appropriate informational base. There may be cases where the claim and the informational base are specified in the same way (e.g. (equal) utility as a claim and utility as the informational base). So, the differentiation between claims and informational bases does sometimes, but not always, make a difference. (P_​ETMP)

(8)

In areas that are characterized by complexity (such as environmental and sustainability issues), a complex specification of the conceptual elements of justice may be in order, that is, any element might need to be specified in multiple ways (e.g. naming several groups of claim holders and of claim addressees). Put differently: there may be more than one answer to each of the questions outlined in Section 3. At the same time, of course, these specifications still need to be consistent with each other. (P_​ETMP)

Hedges are also often used, especially in the philosophical articles, to express caution when making assumptions about other researchers’ interpretations, beliefs and assumptions (9)–​(10): (9)

Of course, this inclusive and dynamic understanding inevitably implies grey areas, which may be interpreted as rendering the framework imprecise. (L_​ AL) (10) A “pure” relationalist might claim we have agent-​relative obligations to not eat a dependent pig. (P_​ETMP)

The corpora include very few examples in which the epistemic property of possibility clearly predominates. The typical examples of this type refer to beliefs that are commonly held in the disciplinary community, and that are not only the author’s beliefs. The corpora suggest that these uses are more common in philosophical research articles than in linguistic ones. Further study is required to address this question. Example (11) is a part of a passage in which the author discusses views of thinking about perception. He assumes that there are two views that are commonly shared in philosophy, and (11) presents the first view. He uses may and may not to show that he thinks it is a matter of dispute whether the content of perceptual states is different from the content of beliefs. May and may not thus express possibility. The author presents two possible interpretations, yet these hedges at the

The Use of English Modal Verbs in Research Articles

371

same time also convey caution and allow the author to negotiate these interpretations with readers. Example (12) further illustrates this kind of usage. The author states what he thinks are two possible positions on the issue he is dealing with, and he presents these positions as possibilities. Both uses of might also involve caution, allowing space for readers to participate in a dialogue. (11)

There are two very different views of thinking about perception. The first one is this. Perceptual states are representations. They represent the world as being a certain way. They have content, which may or may not be different from the content of beliefs. (P_​EJP)

(12)

There are different positions on the epistemological status of considered judgments (or intuitions): a) they might be seen as referring to some ultimate moral truth existing in the world, or b) they might be seen as the result of practical experience in making judgment. (P_​ETMP)

Hedging is also an effective way to mitigate criticisms or differing views. Hedges may be used in these cases to admit a possible interpretation and to show deference to readers. The possible interpretation is then often rejected in favor of a better interpretation, or problems with the possible interpretation are shown. Contrastive markers (e.g. but, however, nevertheless) are often used in these kinds of argumentative contexts. Regarding face, the criticisms or differing views are risky because they threaten the reader’s face, the desire to be approved as a competent researcher. Hedges help here to protect the reader’s face in this respect. I use the term ‘concessive’ in this study for these kinds of usages that are clearly more common in the philosophical corpus. Contrastive devices are also marked in bold in (13)–​(15). For example, the authors in (13) first concede that their study supports one of another researcher’s conclusions. They then state that they would, however, modify the researcher’s conclusions. They propose that her finding may only hold true in restricted circumstances. May is used in this example to mitigate the authors’ own view and to show deference to the researcher whose conclusions they consider. May thus helps to protect this researcher’s face. This kind of use also protects the authors’ face because it helps the authors to present themselves as persons who give respect for colleagues’ views. It should be noted that the concessive uses

372

Liisa Vilkki

of may and might can only be found in complex argumentative contexts that often consist of several sentences. The concessive uses are clearly polite. Mitigation of criticism and deference to readers are predominant properties, although these uses also convey epistemic possibility. (13)

The present study supports her conclusion that because bootstrapping resamples with replacement from the originally observed values, bootstrapped analyses based on very small samples are similarly limited in their ability to represent the population … We would modify Lansing’s conclusions, though, proposing that this finding may only hold true when the effect sizes are larger and therefore more easily detected in tests of statistical significance. (L_​A L)

(14)

Shakespeare’s use of Williams’ claim is unpersuasive: biomedical impairment might be a central part of the experience of disability in some cases but not others. But read as a sort of exhortation, it may have more force: it may be that … (P_​ETMP)

(15)

Aristotle’s argument ‘ from the point of view of nature’ suggests that friendship is a good, in general, because it enables one to perceive that one is alive. The dying man is indeed alive, for now, and Aristotle may be right that the presence of his friends somehow enables him to perceive as much. That perception may indeed be good. But the perception of being alive does not seem apt, in itself, to make ceasing to be alive any better. The value of friendship that Aristotle is elaborating in this second argument therefore seems ill-​suited to explain the particular value that we are hoping to explain. (P_​IJPS)

The modal verbs may and might may also be used to modalize other researchers’ claims (16)–​(17). These kinds of uses are descriptive, and they do not represent the author’s own views (cf. Nuyts 2001: 39–​41). Depending on the context, the descriptive uses could also be interpreted as hedges, if they belong to the author’s hedging strategy. However, I did not find such examples in the two corpora. (16) With the help of the microscope analogy, Hult (2010: 14) explicates that we may focus our object of study and its features depending on how we set the lens. (L_​A L) (17) Now, supposing all this is so, what makes the question of hope central to CPoJ? Well, in Logic, Kant says that the field of philosophy may be summed up in the following questions: What can I know? What I may hope? What ought I to do? What is man? (P_​ETMP)

The Use of English Modal Verbs in Research Articles

373

Hedges and boosters often occur in clusters. The clustering of hedges and boosters is more usual in the philosophical corpus: (18) The Shakesperian view implies that … This may well have benefits, if it helps keep the focus on the complex interaction of genuine impairment and unjust response that generates disability disadvantage in many, maybe most, cases. But it is hardly obvious that these benefits outweigh the cost. (P_​ETMP) (19) Less extreme counterexamples might also be developed. A school bully might be said to pay full attention to their victim while bullying them, a stalker might be said to pay full attention to their victim while harassing them, a bank robber might be said to pay full attention to the safe cracking it, and although we might (but maybe should not, as I will argue in a moment) accept that each of these agents should be said to give their full attention to the object toward which their activities are directed, we will probably all agree that the relevant activities are not valuable, which the opponent does in turn take to show that the Core Claim must be false. (P_​EJP)

We saw earlier that several examples in this section can be interpreted in terms of face. They can also be interpreted in terms of the behavioral expectations and social conventions of the two disciplines, linguistics and philosophy. Following Spencer-​Oatey’s (2015) interpre­ tation of politeness, both cautious and concessive uses indicate politeness. However, concessive uses are predominantly polite, whereas for the cautious uses, the notion of predominance of some property seems to be irrelevant. Of course, my corpus-​based interpretation of the polite uses of the selected modal verbs may not be shared by the readers of the articles, included in the corpora. I would suggest that this question involves different types of readers. For example, a person criticized in an article may feel her face threatened to some degree, although the author of the article has mitigated this criticism. Another reader not involved in this debate in any way may consider these critical statements to show polite deference. It is also possible that different readers interpret caution in different ways, depending at least on their knowledge of the subject discussed in an article. Some readers may clearly feel the author’s cautious view, while other readers may simply think that the author is uncertain of the statements presented. It would be interesting to combine the study of readers’ perspectives with the study of corpora. For example, Hyland (1998a, 2005b) has used this kind of

374

Liisa Vilkki

methodology in his studies. However, the investigation of the readers’ interpretation also involves problems, and it is beyond the scope of this paper to discuss these issues. I would suggest that different types of uses of the selected modal verbs can be described by means of meaning potentials and cognitive domains (cf., for example, Langacker 2006, 2008: 44–​54). All lin­­ guistic expressions are associated with meaning potentials. Meaning potential is a continuum of meanings, the union of individually or collectively remembered uses. Meaning potentials connect an expression both to encyclopedic information and to more specifically linguistic information concerning its use. All types of relevant information are represented by cognitive domains, including pertinent background knowledge concerning, for example, the expression of degrees of certainty, types of inference and types of politeness in research articles of different disciplines. Based on the qualitative analysis, it is not possible to characterize the possible differences in the usage of the modal verbs could, must and should in linguistic and philosophical research articles. These markers are mostly used to express root meanings, and only a few examples of epistemic, inferential and polite usages were found. It is possible, however, to state that all these markers are used in both corpora as multifunctional rhetoric strategies. They simultaneously express more than one meaning, and these meanings may be interpreted as properties of usages. Could is used to indicate weak possibility and caution, and neither of these properties predominate. Must conveys both certain inference and involvement with readers. The data suggest that either of these properties may be foregrounded. Should expresses both probable inference and caution, and neither of the properties is foregrounded. In sum, could and should are used in both corpora as hedging strategies, whereas must is used in the corpora as a boosting strategy. These usages are, however, infrequent. It is also possible to use these three markers to simultaneously express epistemic(-​inferential) and root properties. Some caution and modesty may also be involved. This study focuses on epistemic, inferential and polite usages, so these kinds of examples were not considered in detail. The modal verbs may and might are mostly used as different kinds of hedges in both corpora. The most frequent usage simultaneously

The Use of English Modal Verbs in Research Articles

375

involves caution and possibility, and neither of these properties clearly predominates. This kind of usage is very frequent in both corpora. A subtype of this usage involves cautious assumptions about other researchers’ beliefs, interpretations, and assumptions. The epistemic property of possibility predominates, however. This usage occurs more often in the philosophical corpus than in the linguistic corpus. May and might also occur in complex argumentative contexts in which some view is approved or considered possible but is then signaled problematic or rejected. This study calls these usages concessive. I have not distinguished subtypes of this usage, but the analysis shows that may and might may occur in a passage in which a view is partially approved or in a passage in which an approved view is shown as problematic or rejected, possibly in favor of a better interpretation. These markers may also occur in both passages of the text. Concessive usages simultaneously mitigate criticism, show deference, and express possibility, and they are predominantly polite. These usages occur more often in the philosophical corpus than in the linguistic corpus. The modal verbs may and might may also modalize other researchers’ statements, and they do not represent hedging strategies insofar as these uses are purely descriptive. I did not find examples in which the descriptive uses form part of the author’s hedging strategy.

5. Conclusions I have presented both a quantitative and a qualitative analysis in this paper of the usages of five epistemic(-inferential) modal verbs in research articles in the linguistic corpus and in the philosophical corpus. The detailed contextual analysis shows that may and might are often multifunctional. They are often used in hedging strategies in which they simultaneously convey different kinds of properties. However, could, must and should seem to be used mostly as expressions of root modality. Only a few examples of hedging (could, should) and boosting (must) strategies were found. My first research question was, “What kind of uses of the selected modal verbs can be distinguished

376

Liisa Vilkki

in linguistic and philosophical research articles?”. Section 4 answered this question in detail, and the following summary gives the uses found in the corpora: • • • • •

could: root, root hypothetical, epistemic weak possibility, caution may: root, epistemic possibility, quasi-​subjunctive, caution, concessive, descriptive might: root, root hypothetical, epistemic possibility, epistemic hypothetical, quasi-​subjunctive, caution, concessive, descriptive must: root, inference and epistemic certainty, involvement with the reader should: root, inference and epistemic probability, caution

My purpose was also to explore whether the selected modal verbs are used differently in linguistic and philosophical research articles. Section 4’s qualitative analysis showed first that philosophical articles especially provide plenty of examples in which may and might express caution when making assumptions about other researchers’ interpretations, beliefs and assumptions. The analysis also suggests that philosophical articles contain more usages of may and might that refer to beliefs that are commonly held in the disciplinary community. The clearest difference in the use of may and might was that the philosophical corpus includes more examples of concessive usages of these markers. It was also noted that both hedges and boosters often occur in clusters and that the clustering is more usual in the philosophical corpus. My third research question concerned the possibility of explaining the differences. The nature of philosophical research can probably explain these differences: It mainly involves re-​assessing prior positions and presumptions in the process of dissecting research problems into their parts (cf. Chudnoff 2007, Section 2). Philosophers must take special care of controlling the level of personality to build convincing arguments. The central concepts in philosophical writing are clarity and precision, and one of the features of these concepts may be the avoidance of language expressing personal interests and commitments. This feature may also partly explain why philosophers so often use hedging strategies when referring to colleagues’ views. Linguists also

The Use of English Modal Verbs in Research Articles

377

consider previous research in their studies, but this re-​assessing is usually not in the focus of their interest. That is, the differences in the use of may and might in the two types of research articles could be explained by differences in the conventions and practices of these two disciplines. This study’s findings seem to reflect some disciplinary differences in the usage of especially may and might, yet other factors such as the nature of the subject matter and journal policies may also influence the use of epistemic and inferential modal verbs in linguistic and philosophical research articles. There is also the question of individual differences in the usage of the modal verbs. This question could be discussed in terms of voice and identity (cf. Sections 1 and 2). The present study has not considered the distribution of modal verbs over different sections of the research articles. This is also an important issue for future research. The different kinds of hedging and boosting strategies distinguished in this study could also be studied further in more detail, ideally on bigger corpora.

References Aikhenvald, Alexandra 2006. Evidentiality. Oxford: Oxford University Press. Bargiela-​ Chiappini, Francesca /​Haugh, Michael (eds) 2009. Face, Communication and Social Interaction. London-​Oakville: Equinox. Barton, Ellen L. 1993. Evidentials, argumentation, and epistemological stance. College English 55/​7, 745–​769. Bondi, Marina 2012. Voice in textbooks: between exposition and argument. In Hyland, Ken /​Sancho Guinda, Carmen (eds) Stance and Voice in Written Academic Genres, Basingstoke-​New York: Palgrave Macmillan, 101–​115. Boye, Kasper 2012. Epistemic Meaning: A Crosslinguistic and Functional-​Cognitive Study. Empirical Approaches to Language Typology 43. Berlin-​New York: de Gruyter Mouton.

378

Liisa Vilkki

Bybee, Joan L. /​Perkins, Revere /​Pagliuca, William 1994. The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. Chicago: The University of Chicago Press. Chan, Thomas Hon- Tung 2015. A corpus- based study of the expression of stance in dissertation acknowledgements. Journal of English for Academic Purposes 20, 176–191. Charles, Maggie 2006. The construction of stance in reporting clauses: a cross- disciplinary study of theses. Applied Linguistics 27, 492–518. Chudnoff, Elijah 2007. A Guide to Philosophical Writing. http://​isi​tes. harv​ard.edu/​k 241 Coates, Jennifer 1983. The Semantics of the Modal Auxiliaries. London-​ Canberra: Croom Helm. Crompton, Peter 1997. Hedging in academic writing: some theoretical problems. English for Specific Purposes 16/​4, 271–​287. Dendale, Patrick /​Tasmowski, Liliane 2001. Introduction: evidentiality and related notions. Journal of Pragmatics 33, 339–​348. Diani, Giuliana 2014. Politeness. In Aijmer, Karin /​Rühlemann, Christoph (eds) Corpus Pragmatics: A Handbook, Cambridge: Cambridge University Press, 169–​191. Fløttum, Kjersti 2006. Introduction. In Fløttum, Kjersti /​Gedde-​Dahl, Trine /​Kinn, Torodd (eds) Academic Voices: Across Languages and Disciplines. Pragmatics & Beyond 148, Amsterdam-​ Philadelphia: John Benjamins Publishing Company, 1–​28. Fraser, Bruce 2010. Pragmatic competence: the case of hedging. In Kaltenböck, Gunther /​Mihatsch, Wiltrud /​Schneider, Stefan (eds.) 2010b, New Approaches to Hedging. Studies in Pragmatics 9. Bingley: Emerald, 5–​34. Gross, Alan G. /​Chesley, Paula 2012. Hedging, stance and voice in medical research articles. In Hyland, Ken /​Sancho Guinda, Carmen (eds.) Stance and Voice in Written Academic Genres, Basingstoke-​New York: Palgrave Macmillan, 85–​100. Hartzell, Kai 2018. C+​+​program for the analysis of frequencies and concordances. Manuscript. Heine, Bernd 1995. Agent-​oriented vs. epistemic modality: some observations on German modals. In Bybee, Joan L. /​Fleischman, Suzanne (eds.) Modality in Grammar and Discourse. Typological

The Use of English Modal Verbs in Research Articles

379

Studies in Language 12, Amsterdam-​Philadelphia: John Benjamins Publishing Company, 17–​53. Hinkel, Eli 2005. Hedging, inflating, and persuading in L2 academic writing. Applied Language Learning 15, 29–​53. Hobbs, Valerie 2013. Accounting for the great divide: Features of clarity in analytic philosophy journal articles. Journal of English for Academic Purposes 15, 27–​36. Hunston, Susan 1994. Evaluation and organization in a sample of writtben academic discourse. In Coulthard, Malcolm (ed.) Advances in Written Text Analysis, London: Routledge, 191–​218. Hyland, Ken 1998a. Boosting, hedging and the negotiation of academic knowledge. Text 18/​3, 349–​382. Hyland, Ken 1998b. Hedging in Scientific Research Articles. Amsterdam-​Philadelphia: John Benjamins Publishing Company. Hyland, Ken 2004. Disciplinary Discourses. Social Interactions in Academic Writing. Ann Arbor: The University of Michigan Press. Hyland, Ken 2005a. Metadiscourse: Exploring Interaction in Writing. London: Continuum. Hyland, Ken 2005b. Stance and engagement: a model of interaction in academic discourse. Discourse Studies 7/​2, 173–​192. Hyland, Ken 2015a. Genre, discipline and identity Journal of English for Academic Purposes 19, 32–​43. Hyland, Ken 2015b. Metadiscourse. In Tracy, Karen /​Sandel, Todd /​ Ilie, Cornelia (eds) The International Encyclopedia of Language and Social Interaction, Hoboken, NJ: Wiley. https://​doi.org/​ 10.1002/​978111​8611​463.wbi​elsi​003. Kaltenböck, Gunther /​Mihatsch, Wiltrud /​Schneider, Stefan 2010a. Introduction. In Kaltenböck, Gunther /​Mihatsch, Wiltrud /​ Schneider, Stefan (eds) New Approaches to Hedging. Studies in Pragmatics 9, Bingley: Emerald, 1–​13. Kaltenböck, Gunther /​Mihatsch, Wiltrud /​Schneider, Stefan (eds) 2010b. New Approaches to Hedging. Studies in Pragmatics 9. Bingley: Emerald. Langacker, Ronald 2006. On the continuous debate about discreteness. Cognitive Linguistics 17, 107–​151. Langacker, Ronald W. 2008. Cognitive Grammar: A Basic Introduction. Oxford-​New York: Oxford University Press.

380

Liisa Vilkki

Lave, Jean /​Wenger, Etienne 1991. Situated Learning: Legitimate Peripheral Participation. Cambridge: Cambridge University Press. Mauranen, Anna 1997. Hedging in language revisers’ hands. In Markkanen, Raija /​Schröder, Helmut (eds) Hedging and Discourse. Approaches to the Analysis of a Pragmatic Phenomenon in Academic Texts, Berlin: de Gruyter Walter, 115–​134. Mauranen, Anna 2012. Exploring ELF. Academic English Shaped by Non-​Native Speakers. Cambridge: Cambridge University Press. Myers, Greg 1989. The pragmatics of politeness in scientific articles. Applied Linguistics 10, 1–​35. Nuyts, Jan 2001. Epistemic Modality, Language, and Conceptualization. A Cognitive-​ Pragmatic Perspective. Human Cognitive Processing, Vol. 5, Amsterdam-​Philadelphia: John Benjamins Publishing Company. O’Driscoll, Jim 2017. Face and (im)politeness. In Culpeper, Jonathan /​ Haugh, Michael /​Kádár, Dániel Z. (eds) The Palgrave Handbook of Linguistic (Im)politeness, London: Springer Nature, 89–​118. Oliver, Sonia 2015. Spanish authors dealing with hedging or the challenges of scholarly publication in English L2. In Plo-​Alastrué, Ramón /​Pérez-​Llantada, Carmen (eds) English as a Scientific and Research Language: Debates and Discourses, Berlin-​ Boston: De Gruyter, 141–​157. Palmer, Frank R. 2001. Mood and Modality. 2nd ed. Cambridge: Cambridge University Press. Salager-​Meyer, Françoise 1994. Hedges and textual communicative function in medical English written discourse. English for Specific Purposes 13, 149–​170. Spencer-​Oatey, Helen 2005. (Im)Politeness, face and perceptions of rapport: unpackaging their bases and interrelationships. Journal of Politeness Research 1, 95–​119. Spencer-​Oatey, Helen 2007. Theories of identity and the analysis of face. Journal of Pragmatics 39, 639–​656. Spencer-​Oatey, Helen 2015. Politeness strategies: rapport theory. In Bennett, Janet M (ed.) The SAGE Encyclopedia of Intercultural Competence, Thousand Oaks-​London-​New Delhi: SAGE Publications, 675–​687.

The Use of English Modal Verbs in Research Articles

381

Takimoto, Masahiro 2015. A corpus-​based analysis of hedges and boosters in English academic articles. Indonesian Journal of Applied Linguistics 5/​1, 95–​105. Tardy, Christine M. 2012. Current conceptions of voice. In Hyland, Ken /​Sancho Guinda, Carmen (eds) Stance and Voice in Written Academic Genres, Basingstoke-​New York: Palgrave Macmillan, 35–​ 48. Terkourafi, Marina 2015. Conventionalization: a new agenda for im/​ politeness research. Journal of Pragmatics 86, 11–​18. Terkourafi, Marina /​Kádár, Dániel Z. 2017. Convention and ritual (im) politeness. In Culpeper, Jonathan /​Haugh, Michael /​Kádár, Dániel Z. (eds) The Palgrave Handbook of Linguistic (Im)politeness, London: Springer Nature, 171–​195. Tse, Polly /​Hyland, Ken 2006. Gender and discipline: exploring metadiscourse variation in academic book reviews. In Hyland, Ken /​Bondi, Marina (eds) Academic discourse across disciplines, Bern: Peter Lang, 2006, 177–​202. Vilkki, Liisa 2006. Politeness, face and facework: current issues. In Suominen, Mickael & al. (eds) A Man of Measure: Festschrift in Honour of Fred Karlsson on his 60th Birthday. A Special Supplement to SKY Journal of Linguistics 19, Turku: The Linguistic Association of Finland, 322–​332. [available at http://​www.ling. helsi​nki.fi/​sky/​julkai​sut/​sky​2006​spec​ial.shtml]. Vilkki, Liisa forthcoming. Semantic Typology of Grammatically Expressed Epistemic Modality and Inferentiality. Vold, Eva Thue 2006. Epistemic modality markers in linguistic and medical research articles. In Hyland, Ken /​Bondi, Marina (eds) 2006 Academic discourse across disciplines, Bern: Peter Lang, 225–​249. Wang, Jiayi /​Spencer-​Oatey, Helen 2015. The gains and losses of face in ongoing intercultural interaction: a case study of Chinese participant perspectives. Journal of Pragmatics 89, 50–​65. Watts, Richard 2003. Politeness. Cambridge: Cambridge University Press. Xie, Jianping 2020. A review of research on authorial evaluation in English academic writing: A methodological perspective. Journal of English for Academic Purposes 47, 1–​20.

Notes on Contributors

Aoife Kathleen AHERN is Associate Professor at the Universidad Complutense de Madrid School of Education, where she works in the field of language teacher education. Her research interests are related to the interfaces of grammar and pragmatics, as well as to second language teaching, learning and acquisition, focussed on English and Spanish, and to literacy education. She has published several books such as Pragmática, as co-​author with V. Escandell-​Vidal and J. Amenós Pons (Madrid: Akal 2020); Comunicación y cognición en ELE: La perspectiva pragmática, co-​authored with J. Amenós Pons and V. Escandell-​ Vidal; Procedural Meaning: Problems and Perspectives, as co-​author with V. Escandell-​Vidal and M. Leonetti (Leiden: Brill, 2011). Her research, carried out through several projects funded by the Spanish Ministry for Education and Science and the Ministry for Economy and Competitivity of Spain, has also led to a range of articles in scientific journals and book chapters in collective volumes, on topics in the fields of pragmatics, language and literacy education and second language acquisition. José AMENÓS-​PONS is a Senior Reader in Spanish Linguistics at the Universidad Complutense de Madrid. He specializes in cognitive pragmatics and foreign language acquisition processes and has published numerous articles on tense, aspect and mood, and more generally, on the relationships between grammar and pragmatics, and on how this relationship influences L2 learning. He has participated in different research projects funded by the Spanish national research authority, almost uninterruptedly from 2007 to the present day: Procedural Semantics and Explicit Content 1, 2 and 3 (2007–​2015), The Semantics/​ Pragmatics Interface and the Resolution of Interpretive Mismatches (2015–​2019), Evidentiality, Perspectivisation and Subjectivisation at the Interfaces of Language (2020–​2023). He has taught Spanish as a second language for many years and has been very actively involved

384

Notes on Contributors

in L2 teacher training. He has co-​authored L2 Spanish textbooks and materials for teacher education. Patrick DENDALE is professor of French Language and French Linguistics working at the University of Antwerp. With his PhD (1991) on evidential marking in French (Le marquage épistémique de l’énoncé: esquisse d’une théorie avec applications au français), he was the first scholar to propose evidential analyses of French items, previously analyzed as modal items (the reportive conditional and epistemic devoir). He has published a bibliography on French evidential markers (2014) and a whole series of studies on a variety of French grammatical and lexical items with an evidential function (the conjectural future, the conjectural conditional, visiblement, certainement, à ce que je vois, à vue d’œil, à vue de nez, selon, aux dires de, quelque chose me dit que) and a long overview article on evidentials in French (Dendale 2022). He has published also some well cited studies on evidentiality in general (Dendale & Tasmowski 2001) and two thematic issues on the topic, one on French evidential markers (Langue française 1994) and one on evidentiality and related notions (Journal of Pragmatics 2001), and further collective volumes on modal verbs (2001) and on commitment (2008, 2009, 2011). Most recently he has focused his attention on the problem of the very nature of inferential evidentiality and of "evidential inference" and the way this problem is treated in the literature. He teaches a MA course on evidential and epistemic marking in French (since 2010). Oana-​Adriana DUȚĂ teaches Spanish linguistics and translation at the University of Craiova, Romania. She holds a PhD from the University of Bucharest with a dissertation on Spanish and Romanian phraseology and has published three Spanish language coursebooks and over 60 articles in scientific journals. She has delivered presentations on contrastive linguistics and translation in national and international conferences in Spain, Italy, France, Poland, the Czech Republic and Romania. She has performed extensive research on foreign language teaching in 6 international projects focusing on education within the Erasmus KA2 and the EEA Grants programs and has co-​edited 4 volumes, such as, inter alia, Sustainable and Solidary Education. Reflections and Practices (Peter Lang, 2017). She has translated 14 books for prestigious publishing houses in Romania, such as Litera and RAO,

Notes on Contributors

385

and has been working as an external translator for European Union institutions for more than ten years. She is the editorial secretary of the Scopus-​indexed journal Annals of the University of Craiova. Philology. Linguistics and is constantly focused on improving her lifelong training –​as she holds a PhD in Economics and a postgraduate diploma in terminology from the Pompeu Fabra University of Barcelona. Her research interests include contrastive linguistics, cognitive semantics and translation. Pedro GUIJARRO-​FUENTES is a Professor in Spanish linguistics at the University of the Balearic Islands and has headed multiple projects on the acquisition of various phenomena on Spanish language including, but not limited to, the acquisition of ser/​estar, tense and mood, differential object marking, all funded by the British Academy, Art and Humanities Research Council, the Deutsche Forschungsgemeinschaft, Spanish Ministry of Economy and Competitiveness and more recently by the European Union research foundation under the Horizon 2020. He has authored and co-​authored numerous journal articles, book chapters and books, published by various international publishers and in high-​impact journals. He has organized many professional meetings and has been a participant in many professional conferences. Andra KALNAČA is a Professor and Senior Researcher of Latvian Linguistics at the University of Latvia (Rīga). Her research interests relate to morphology, morphosyntax and morphophonology, functional and cognitive grammar, modality, and evidentiality. Her core work is focused on Latvian grammar and the theory of grammar. She is the author of A Typological Perspective on Latvian Grammar (2014) and Latvian Grammar (2021; with Ilze Lokmane). Dorota KOTWICA is an Assistant Professor of Spanish at the University of Valencia, from which she received her PhD in Linguistics in 2016. Her postdoctoral research project (2019–​2021) focused on the expression of common knowledge and folklore evidentiality in Spanish and its pragmatic-​discursive functions. She is a member of Val.Es.Co. research group, dedicated to the study of oral language and colloquial register. Her research interests include discourse analysis, evidentiality and pragmatic categories of mitigation and intensification. She

386

Notes on Contributors

published the monograph La evidencialidad en el artículo científico. Historia de un género discursivo de 1799 a 1920 (Peter Lang, 2019), as well as several journal articles and book chapters. Ilze LOKMANE is an Associate Professor and Senior Researcher of Latvian Linguistics at the University of Latvia (Rīga). Her research interests relate to syntax, morphosyntax, pragmatics, functional and cognitive grammar, modality and evidentiality. Her core work is focused on Latvian grammar and the theory of pragmatics. She is the author of Latvian Grammar (2021; with Andra Kalnača). Johanna MIECZNIKOWSKI studied Italian Philology, Russian Philology, French Linguistics and General Linguistics at the universities of Basel and Amsterdam. In 2002, she obtained a PhD in French Linguistics at Basel University. Her research in that period concentrated on biographies of multilingual speakers and on the interactive construction of scientific discourse in a multilingual environment. After a stay at the University of Turin, where she investigated the conditional form and discourse markers in spoken Italian and French, Basel University granted her, in 2010, the Habilitation to teach Romance Linguistics. Since 2007 she has been affiliated to USI Università della Svizzera italiana, Lugano, where she teaches Linguistics and Second Language Teaching and Learning within the Bachelor's and Master's programs in Italian language, literature and culture. Her recent research has regarded modality and argumentation (Swiss National Science Foundation grant no. 120740), inferential uses of perception predicates (SNSF grant no. 141350), the categorization of information sources in spoken Italian (SNSF grant no. 192771, ongoing), and teaching materials for Italian as a foreign language. From 2012 to 2016, she was chair of the Swiss Association for Applied Linguistics. Natalia MORA-​LÓPEZ graduated in English and Spanish Studies at the Complutense University of Madrid. She also completed a master’s degree in English Linguistics and in Teacher Training in Secondary Education, Vocational Training and Languages. She obtained her PhD in English Linguistics at the same university. Regarding her professional career, Natalia has worked as a computational linguist, translator and proofreader on the one hand, and as a teacher of English and

Notes on Contributors

387

Spanish language and literature on the other. She has taught at several universities like Valencian International University, International University of La Rioja, CEU San Pablo University and Nebrija University. She currently works at Complutense University of Madrid, where she teaches English for Specific Purposes. Her research interests focus on the use of evaluative language and the expression of opinion, digital media, and teaching innovation. She has participated in national and international conferences, and her research has been published by prestigious publishing houses like John Benjamins and Peter Lang. She has also participated in national and European research projects related to both linguistics and teaching innovation. Tanja MORTELMANS is Professor of German Language and Linguistics at the University of Antwerp (Belgium). Her main research interests are the expression of modality (by means of modal verbs and mood markers), evidentiality (with a focus on reportative and inferential evidentiality), progressive aspect and future tense in German, Dutch and English, mainly from a cognitive-​functional and/​or contrastive perspective. Furthermore, grammaticalization phenomena, (inter) subjectification and grounding are recurrent topics in her research. Her publications include a monograph on functional linguistic theories (together with Elena Smirnova, De Gruyter 2010) and a number of research articles in collective volumes and international journals like Journal of Pragmatics, Linguistics and Journal of Germanic linguistics. She is co-​editor of the issue of the Belgian Journal of Linguistics (BJL 34) on Construction Grammar (John Benjamins, 2020), and of the books Aspects of Linguistic Variation (John Benjamins, 2018), Marqueurs temporels et modaux en usage (Rodopi, 2013) and Grammaticalization and Language Change (John Benjamins, 2012). She is a member of the editorial board of the Belgian Journal of Linguistics (John Benjamins) and of Germanistische Mitteilungen (Winter). At the University of Antwerp, she lectures in German grammar, grammaticalization, modality and gender. Cecilia Mihaela POPESCU is a Professor in Applied Romance Linguistics, General Linguistics, and Pragmatics at the University of Craiova, Romania. She received her PhD in 2006, in the field of Philology (Romance Linguistics subdomain), at the same academic institution,

388

Notes on Contributors

under the supervision of Prof. Maria Iliescu. In her dissertation, she carried out a comparative study on the expression of the potential and the irrealis in Latin, French, and Romanian. Actually, her research activity is structured around four major topics, which are somehow complementary to each other and which relate to: (a) furthering the studies on Latin linguistics issues, (b) studying the semantic-​functional behavior of certain verbal compartments (future, subjunctive, conditional, presumptive) in Romance languages viewed in synchronicity, diachrony, but also in contrast, (c) analyzing lexical borrowings from French to Romanian, and (d) studying discursive markers in Romance languages. In 2013, she published her postdoctoral thesis entitled The Future and the Conditional in Romance Languages. A Morphosyntactic and Semantic Approach from a Diachronic Perspective. She also published papers in Revue roumaine de linguistique, Studii şi cercetări lingvistice, Revue de Linguistique Romane, Zeitschrift fur Franzosische Sprache und Literatur, Revista de Filología Románica, Revue de Sémantique et Pragmatique, Discours. Revue de linguistique, psycholinguistique et informatique etc, as well as other international peer-​ reviewed volumes. She has been visiting scholar at Charles de Gaulle University, Lille3 (2002), Denis Diderot University, Paris 7 (2009), and University of Rome 3 (2011). She was the director of an advanced research project funded internally, and member of several national grants and international projects (EEA grants, POSDRU or AUF projects). Cecilia-​Mihaela Popescu is co-​organizer of national and international conferences on historical linguistics and lexicology and has edited various collective volumes. From 2016 up to 2022, she was selected member in the Board of the International Society of Romance Linguistics (Société de Linguistique romane). Tabea REINER is currently researching and teaching at LMU Munich in the department of German Linguistics. Her research interests include (but are not limited to) linguistic theories and grammatical semantics, focussing on the verbal domain. In particular, she has been working on periphrasis: do function words (including auxiliaries) belong to the domain of grammar, do periphrases belong to syntax or morphology, how productive are certain periphrases with a given semantics, e.g. can (periphrastic) posterior infinitives be built from future periphrases?

Notes on Contributors

389

With respect to languages considered, she mostly works on German but from a General Linguistics perspective. Andreu SENTÍ is a tenured Associate Professor of Catalan linguistics at the University of València. He received his PhD in Catalan linguistics from the University of Alacant with an extraordinary mention in 2013. His research interests range from semantic change, historical linguistics, grammaticalization, dialectology and language learning. He has authored numerous works on modal and evidential constructions and grammaticalization in Catalan. He is currently coordinating the dialectological fieldwork and design of the Corpus Parlars, a spoken and informal corpus for Valencian Catalan, and running the research group VaLingCat (Linguistic Variation in Catalan). Dr. Sentí has taught courses of Catalan linguistics, historical linguistics, applied linguistics, writing skills and stylistics and public speaking at the University of València. He has been visiting scholar at KULeuven, Ghent University, University of Barcelona, University of Cambridge, University of Birmingham and Pompeu Fabra University. Nicolas TOURNADRE is Emeritus Professor of linguistics at Aix-​ Marseille University and member of the University Institute of France (IUF). He conducts research in the field of linguistic typology and has undertaken extensive fieldwork in Asia, particularly in China, India, Nepal, Bhutan and Tajikistan. His research interests include the study of the syntax-​semantics interface as well as ergativity and more recently TAM and evidentiality. N. Tournadre is a specialist of the Tibetic languages but he has also worked on languages belonging to other families such as Indo-​Iranian, Slavic and Semitic. His major books include L’ergativité en tibétain moderne [ergativity in Modern Tibetan] (Peeters, 1996); Le Prisme des langues [the Prism of Languages] (Asiathèque, 2014); Sherpa-​English and English-​Sherpa Dictionary: With Literary Tibetan and Nepali Equivalents [with Lhakpa Norbu Sherpa, Gyurme Chodrak and Guillaume Oisel] (Vajra Publications, 2009); The Manual of Standard Tibetan [with Sangda Dorje] (Snowlion, 2003); The Tibetic Languages, an Introduction to the Family of Languages derived from Old Tibetan [with Hiroyuki Suzuki] (CNRS, Lacito publications, 2022); Linguistic Panorama of Tadjikistan –​Crossroad of Indo-​European languages: Iranic, Indo-​aryan and Slavic [with H. Lessan-​Pezechki and

390

Notes on Contributors

C. Aslanov (eds.)] (Presse Universitaire de Provence, forthcoming). He has given many presentations and talks in Asia (China, India, Bhutan, Myanmar, Singapore, Tajikistan, Japan), in Russia, Northern America (USA, Canada) and Southern America (Brazil), in the Middle East (Israel, Iran), in Africa (Morocco), in the Caucasus (Georgia), in Australia and in most European countries. Liisa VILKKI’s study has mainly been in the fields of epistemic modality, evidentiality, semantic typology, and linguistic politeness and facework. She has taught and held positions in several disciplines at the University of Helsinki, Finland. Earlier, she also worked on computational morphology in the private sector. She is currently working on her PhD thesis in general linguistics at the University of Helsinki. In the thesis, she proposes a semantic typology of grammatically expressed epistemic modality and inferentiality. The study is based on a sample of 110 languages of the world. Currently, she also runs her own training company, focusing on written communication.

Notes on Contributors

391

Linguistic Insights Studies in Language and Communication

This series aims to promote specialist language studies in the fields of linguistic theory and applied linguistics, by publishing volumes that focus on specific aspects of language use in one or several languages and provide valuable insights into language and communication research. A cross-disciplinary approach is favoured and most European languages are accepted. The series includes two types of books: – Monographs – featuring in-depth studies on special aspects of language theory, language analysis or language teaching. – Collected papers – assembling papers from workshops, conferences or symposia. Each volume of the series is subjected to a double peer-reviewing process. Vol.

1

Maurizio Gotti & Marina Dossena (eds) Modality in Specialized Texts. Selected Papers of the 1st CERLIS Conference. 421 pages. 2001. ISBN 3-906767-10-8 · US-ISBN 0-8204-5340-4

Vol.

2

Giuseppina Cortese & Philip Riley (eds) Domain-specific English. Textual Practices across Communities and Classrooms. 420 pages. 2002. ISBN 3-906768-98-8 · US-ISBN 0-8204-5884-8

Vol.

3

Maurizio Gotti, Dorothee Heller & Marina Dossena (eds) Conflict and Negotiation in Specialized Texts. Selected Papers of the 2nd CERLIS Conference. 470 pages. 2002. ISBN 3-906769-12-7 · US-ISBN 0-8204-5887-2

Vol.

4

Maurizio Gotti, Marina Dossena, Richard Dury, Roberta Facchinetti & Maria Lima Variation in Central Modals. A Repertoire of Forms and Types of Usage in Middle English and Early Modern English. 364 pages. 2002. ISBN 3-906769-84-4 · US-ISBN 0-8204-5898-8

Editorial address: Prof. Maurizio Gotti

Università di Bergamo, Dipartimento di Lingue, Letterature e Culture Straniere Piazza Rosate 2, 24129 Bergamo, Italy Fax: +39 035 2052789, E-Mail: [email protected]

392

Notes on Contributors

Vol.

5

Stefania Nuccorini (ed.) Phrases and Phraseology. Data and Descriptions. 187 pages. 2002. ISBN 3-906770-08-7 · US-ISBN 0-8204-5933-X

Vol.

6

Vijay Bhatia, Christopher N. Candlin & Maurizio Gotti (eds) Legal Discourse in Multilingual and Multicultural Contexts. Arbitration Texts in Europe. 385 pages. 2003. ISBN 3-906770-85-0 · US-ISBN 0-8204-6254-3

Vol.

7

Marina Dossena & Charles Jones (eds) Insights into Late Modern English. 2nd edition. 378 pages. 2003, 2007. ISBN 978-3-03911-257-9 · US-ISBN 978-0-8204-8927-8

Vol.

8

Maurizio Gotti Specialized Discourse. Linguistic Features and Changing Conventions. 351 pages. 2003, 2005. ISBN 3-03910-606-6 · US-ISBN 0-8204-7000-7

Vol.

9

Alan Partington, John Morley & Louann Haarman (eds) Corpora and Discourse. 420 pages. 2004. ISBN 3-03910-026-2 · US-ISBN 0-8204-6262-4

Vol.

10

Martina Möllering The Acquisition of German Modal Particles. A Corpus-Based Approach. 290 pages. 2004. ISBN 3-03910-043-2 · US-ISBN 0-8204-6273-X

Vol.

11

David Hart (ed.) English Modality in Context. Diachronic Perspectives. 261 pages. 2003. ISBN 3-03910-046-7 · US-ISBN 0-8204-6852-5

Vol.

12

Wendy Swanson Modes of Co-reference as an Indicator of Genre. 430 pages. 2003. ISBN 3-03910-052-1 · US-ISBN 0-8204-6855-X

Vol.

13

Gina Poncini Discursive Strategies in Multicultural Business Meetings. 2nd edition. 338 pages. 2004, 2007. ISBN 978-3-03911-296-8 · US-ISBN 978-0-8204-8937-7

Vol.

14

Christopher N. Candlin & Maurizio Gotti (eds) Intercultural Aspects of Specialized Communication. 2nd edition. 369 pages. 2004, 2007. ISBN 978-3-03911-258-6 · US-ISBN 978-0-8204-8926-1

Vol.

15

Gabriella Del Lungo Camiciotti & Elena Tognini Bonelli (eds) Academic Discourse. New Insights into Evaluation. 234 pages. 2004. ISBN 3-03910-353-9 · US-ISBN 0-8204-7016-3

Vol.

16

Marina Dossena & Roger Lass (eds) Methods and Data in English Historical Dialectology. 405 pages. 2004. ISBN 3-03910-362-8 · US-ISBN 0-8204-7018-X

Vol.

17

Judy Noguchi The Science Review Article. An Opportune Genre in the Construction of Science. 274 pages. 2006. ISBN 3-03910-426-8 · US-ISBN 0-8204-7034-1

Notes on Contributors

Vol.

18

Giuseppina Cortese & Anna Duszak (eds) Identity, Community, Discourse. English in Intercultural Settings. 495 pages. 2005. ISBN 3-03910-632-5 · US-ISBN 0-8204-7163-1

Vol.

19

Anna Trosborg & Poul Erik Flyvholm Jørgensen (eds) Business Discourse. Texts and Contexts. 250 pages. 2005. ISBN 3-03910-606-6 · US-ISBN 0-8204-7000-7

Vol.

20

Christopher Williams Tradition and Change in Legal English. Verbal Constructions in Prescriptive Texts. 2nd revised edition. 216 pages. 2005, 2007. ISBN 978-3-03911-444-3.

Vol.

21

Katarzyna Dziubalska-Kolaczyk & Joanna Przedlacka (eds) English Pronunciation Models: A Changing Scene. 2nd edition. 476 pages. 2005, 2008. ISBN 978-3-03911-682-9.

Vol.

22

Christián Abello-Contesse, Rubén Chacón-Beltrán, M. Dolores López-Jiménez & M. Mar Torreblanca-López (eds) Age in L2 Acquisition and Teaching. 214 pages. 2006. ISBN 3-03910-668-6 · US-ISBN 0-8204-7174-7

Vol.

23

Vijay K. Bhatia, Maurizio Gotti, Jan Engberg & Dorothee Heller (eds) Vagueness in Normative Texts. 474 pages. 2005. ISBN 3-03910-653-8 · US-ISBN 0-8204-7169-0

Vol.

24

Paul Gillaerts & Maurizio Gotti (eds) Genre Variation in Business Letters. 2nd printing. 407 pages. 2008. ISBN 978-3-03911-681-2.

Vol.

25

Ana María Hornero, María José Luzón & Silvia Murillo (eds) Corpus Linguistics. Applications for the Study of English. 2nd printing. 526 pages. 2006, 2008. ISBN 978-3-03911-726-0

Vol.

26

J. Lachlan Mackenzie & María de los Ángeles Gómez-González (eds) Studies in Functional Discourse Grammar. 259 pages. 2005. ISBN 3-03910-696-1 · US-ISBN 0-8204-7558-0

Vol.

27

Debbie G. E. Ho Classroom Talk. Exploring the Sociocultural Structure of Formal ESL Learning. 2nd edition. 254 pages. 2006, 2007. ISBN 978-3-03911-434-4

Vol.

28

Javier Pérez-Guerra, Dolores González-Álvarez, Jorge L. Bueno-Alonso & Esperanza Rama-Martínez (eds) ‘Of Varying Language and Opposing Creed’. New Insights into Late Modern English. 455 pages. 2007. ISBN 978-3-03910-788-9

Vol.

29

Francesca Bargiela-Chiappini & Maurizio Gotti (eds) Asian Business Discourse(s). 350 pages. 2005. ISBN 3-03910-804-2 · US-ISBN 0-8204-7574-2

Vol.

30

Nicholas Brownlees (ed.) News Discourse in Early Modern Britain. Selected Papers of CHINED 2004. 300 pages. 2006. ISBN 3-03910-805-0 · US-ISBN 0-8204-8025-8

Vol.

31

Roberta Facchinetti & Matti Rissanen (eds) Corpus-based Studies of Diachronic English. 300 pages. 2006. ISBN 3-03910-851-4 · US-ISBN 0-8204-8040-1

393

394

Notes on Contributors

Vol.

32

Marina Dossena & Susan M. Fitzmaurice (eds) Business and Official Correspondence. Historical Investigations. 209 pages. 2006. ISBN 3-03910-880-8 · US-ISBN 0-8204-8352-4

Vol.

33

Giuliana Garzone & Srikant Sarangi (eds) Discourse, Ideology and Specialized Communication. 494 pages. 2007. ISBN 978-3-03910-888-6

Vol.

34

Giuliana Garzone & Cornelia Ilie (eds) The Use of English in Institutional and Business Settings. An Intercultural Perspective. 372 pages. 2007. ISBN 978-3-03910-889-3

Vol.

35

Vijay K. Bhatia & Maurizio Gotti (eds) Explorations in Specialized Genres. 316 pages. 2006. ISBN 3-03910-995-2 · US-ISBN 0-8204-8372-9

Vol.

36

Heribert Picht (ed.) Modern Approaches to Terminological Theories and Applications. 432 pages. 2006. ISBN 3-03911-156-6 · US-ISBN 0-8204-8380-X

Vol.

37

Anne Wagner & Sophie Cacciaguidi-Fahy (eds) Legal Language and the Search for Clarity / Le langage juridique et la quête de clarté. Practice and Tools / Pratiques et instruments. 487 pages. 2006. ISBN 3-03911-169-8 · US-ISBN 0-8204-8388-5

Vol.

38

Juan Carlos Palmer-Silveira, Miguel F. Ruiz-Garrido & Inmaculada Fortanet-Gómez (eds) Intercultural and International Business Communication. Theory, Research and Teaching. 2nd edition. 343 pages. 2006, 2008. ISBN 978-3-03911-680-5

Vol.

39

Christiane Dalton-Puffer, Dieter Kastovsky, Nikolaus Ritt & Herbert Schendl (eds) Syntax, Style and Grammatical Norms. English from 1500–2000. 250 pages. 2006. ISBN 3-03911-181-7 · US-ISBN 0-8204-8394-X

Vol.

40

Marina Dossena & Irma Taavitsainen (eds) Diachronic Perspectives on Domain-Specific English. 280 pages. 2006. ISBN 3-03910-176-0 · US-ISBN 0-8204-8391-5

Vol.

41

John Flowerdew & Maurizio Gotti (eds) Studies in Specialized Discourse. 293 pages. 2006. ISBN 3-03911-178-7

Vol.

42

Ken Hyland & Marina Bondi (eds) Academic Discourse Across Disciplines. 320 pages. 2006. ISBN 3-03911-183-3 · US-ISBN 0-8204-8396-6

Vol.

43

Paul Gillaerts & Philip Shaw (eds) The Map and the Landscape. Norms and Practices in Genre. 256 pages. 2006. ISBN 3-03911-182-5 · US-ISBN 0-8204-8395-4

Vol.

44

Maurizio Gotti & Davide Giannoni (eds) New Trends in Specialized Discourse Analysis. 301 pages. 2006. ISBN 3-03911-184-1 · US-ISBN 0-8204-8381-8

Vol.

45

Maurizio Gotti & Françoise Salager-Meyer (eds) Advances in Medical Discourse Analysis. Oral and Written Contexts. 492 pages. 2006. ISBN 3-03911-185-X · US-ISBN 0-8204-8382-6

Notes on Contributors

Vol.

46

Maurizio Gotti & Susan Šarcevi´c (eds) Insights into Specialized Translation. 396 pages. 2006. ISBN 3-03911-186-8 · US-ISBN 0-8204-8383-4

Vol.

47

Khurshid Ahmad & Margaret Rogers (eds) Evidence-based LSP. Translation, Text and Terminology. 584 pages. 2007. ISBN 978-3-03911-187-9

Vol.

48

Hao Sun & Dániel Z. Kádár (eds) It’s the Dragon’s Turn. Chinese Institutional Discourses. 262 pages. 2008. ISBN 978-3-03911-175-6

Vol.

49

Cristina Suárez-Gómez Relativization in Early English (950-1250). the Position of Relative Clauses. 149 pages. 2006. ISBN 3-03911-203-1 · US-ISBN 0-8204-8904-2

Vol.

50

Maria Vittoria Calvi & Luisa Chierichetti (eds) Nuevas tendencias en el discurso de especialidad. 319 pages. 2006. ISBN 978-3-03911-261-6

Vol.

51

Mari Carmen Campoy & María José Luzón (eds) Spoken Corpora in Applied Linguistics. 274 pages. 2008. ISBN 978-3-03911-275-3

Vol.

52

Konrad Ehlich & Dorothee Heller (Hrsg.) Die Wissenschaft und ihre Sprachen. 323 pages. 2006. ISBN 978-3-03911-272-2

Vol.

53

Jingyu Zhang The Semantic Salience Hierarchy Model. The L2 Acquisition of Psych Predicates 273 pages. 2007. ISBN 978-3-03911-300-2

Vol.

54

Norman Fairclough, Giuseppina Cortese & Patrizia Ardizzone (eds) Discourse and Contemporary Social Change. 555 pages. 2007. ISBN 978-3-03911-276-0

Vol.

55

Jan Engberg, Marianne Grove Ditlevsen, Peter Kastberg & Martin Stegu (eds) New Directions in LSP Teaching. 331 pages. 2007. ISBN 978-3-03911-433-7

Vol.

56

Dorothee Heller & Konrad Ehlich (Hrsg.) Studien zur Rechtskommunikation. 322 pages. 2007. ISBN 978-3-03911-436-8

Vol.

57

Teruhiro Ishiguro & Kang-kwong Luke (eds) Grammar in Cross-Linguistic Perspective. The Syntax, Semantics, and Pragmatics of Japanese and Chinese. 304 pages. 2012. ISBN 978-3-03911-445-0

Vol.

58

Carmen Frehner Email – SMS – MMS 294 pages. 2008. ISBN 978-3-03911-451-1

Vol.

59

Isabel Balteiro The Directionality of Conversion in English. A Dia-Synchronic Study. 276 pages. 2007. ISBN 978-3-03911-241-8

Vol.

60

Maria Milagros Del Saz Rubio English Discourse Markers of Reformulation. 237 pages. 2007. ISBN 978-3-03911-196-1

395

396

Notes on Contributors

Vol.

61

Sally Burgess & Pedro Martín-Martín (eds) English as an Additional Language in Research Publication and Communication. 259 pages. 2008. ISBN 978-3-03911-462-7

Vol.

62

Sandrine Onillon Pratiques et représentations de l’écrit. 458 pages. 2008. ISBN 978-3-03911-464-1

Vol.

63

Hugo Bowles & Paul Seedhouse (eds) Conversation Analysis and Language for Specific Purposes. 2nd edition. 337 pages. 2007, 2009. ISBN 978-3-0343-0045-2

Vol.

64

Vijay K. Bhatia, Christopher N. Candlin & Paola Evangelisti Allori (eds) Language, Culture and the Law. The Formulation of Legal Concepts across Systems and Cultures. 342 pages. 2008. ISBN 978-3-03911-470-2

Vol.

65

Jonathan Culpeper & Dániel Z. Kádár (eds) Historical (Im)politeness. 300 pages. 2010. ISBN 978-3-03911-496-2

Vol.

66

Linda Lombardo (ed.) Using Corpora to Learn about Language and Discourse. 237 pages. 2009. ISBN 978-3-03911-522-8

Vol.

67

Natsumi Wakamoto Extroversion/Introversion in Foreign Language Learning. Interactions with Learner Strategy Use. 159 pages. 2009. ISBN 978-3-03911-596-9

Vol.

68

Eva Alcón-Soler (ed.) Learning How to Request in an Instructed Language Learning Context. 260 pages. 2008. ISBN 978-3-03911-601-0

Vol.

69

Domenico Pezzini The Translation of Religious Texts in the Middle Ages. 428 pages. 2008. ISBN 978-3-03911-600-3

Vol.

70

Tomoko Tode Effects of Frequency in Classroom Second Language Learning. Quasi-experiment and stimulated-recall analysis. 195 pages. 2008. ISBN 978-3-03911-602-7

Vol.

71

Egor Tsedryk Fusion symétrique et alternances ditransitives. 211 pages. 2009. ISBN 978-3-03911-609-6

Vol.

72

Cynthia J. Kellett Bidoli & Elana Ochse (eds) English in International Deaf Communication. 444 pages. 2008. ISBN 978-3-03911-610-2

Vol.

73

Joan C. Beal, Carmela Nocera & Massimo Sturiale (eds) Perspectives on Prescriptivism. 269 pages. 2008. ISBN 978-3-03911-632-4

Vol.

74

Carol Taylor Torsello, Katherine Ackerley & Erik Castello (eds) Corpora for University Language Teachers. 308 pages. 2008. ISBN 978-3-03911-639-3

Notes on Contributors

Vol.

75

María Luisa Pérez Cañado (ed.) English Language Teaching in the European Credit Transfer System. Facing the Challenge. 251 pages. 2009. ISBN 978-3-03911-654-6

Vol.

76

Marina Dossena & Ingrid Tieken-Boon van Ostade (eds) Studies in Late Modern English Correspondence. Methodology and Data. 291 pages. 2008. ISBN 978-3-03911-658-4

Vol.

77

Ingrid Tieken-Boon van Ostade & Wim van der Wurff (eds) Current Issues in Late Modern English. 436 pages. 2009. ISBN 978-3-03911-660-7

Vol.

78

Marta Navarro Coy (ed.) Practical Approaches to Foreign Language Teaching and Learning. 297 pages. 2009. ISBN 978-3-03911-661-4

Vol.

79

Qing Ma Second Language Vocabulary Acquisition. 333 pages. 2009. ISBN 978-3-03911-666-9

Vol.

80

Martin Solly, Michelangelo Conoscenti & Sandra Campagna (eds) Verbal/Visual Narrative Texts in Higher Education. 384 pages. 2008. ISBN 978-3-03911-672-0

Vol.

81

Meiko Matsumoto From Simple Verbs to Periphrastic Expressions: The Historical Development of Composite Predicates, Phrasal Verbs, and Related Constructions in English. 235 pages. 2008. ISBN 978-3-03911-675-1

Vol.

82

Melinda Dooly Doing Diversity. Teachers’ Construction of Their Classroom Reality. 180 pages. 2009. ISBN 978-3-03911-687-4

Vol.

83

Victoria Guillén-Nieto, Carmen Marimón-Llorca & Chelo Vargas-Sierra (eds) Intercultural Business Communication and Simulation and Gaming Methodology. 392 pages. 2009. ISBN 978-3-03911-688-1

Vol.

84

Maria Grazia Guido English as a Lingua Franca in Cross-cultural Immigration Domains. 285 pages. 2008. ISBN 978-3-03911-689-8

Vol.

85

Erik Castello Text Complexity and Reading Comprehension Tests. 352 pages. 2008. ISBN 978-3-03911-717-8

Vol.

86

Maria-Lluisa Gea-Valor, Isabel García-Izquierdo & Maria-José Esteve (eds) Linguistic and Translation Studies in Scientific Communication. 317 pages. 2010. ISBN 978-3-0343-0069-8

Vol.

87

Carmen Navarro, Rosa Mª Rodríguez Abella, Francesca Dalle Pezze & Renzo Miotti (eds) La comunicación especializada. 355 pages. 2008. ISBN 978-3-03911-733-8

397

398

Notes on Contributors

Vol.

88

Kiriko Sato The Development from Case-Forms to Prepositional Constructions in Old English Prose. 231 pages. 2009. ISBN 978-3-03911-763-5

Vol.

89

Dorothee Heller (Hrsg.) Formulierungsmuster in deutscher und italienischer Fachkommunikation. Intra- und interlinguale Perspektiven. 315 pages. 2008. ISBN 978-3-03911-778-9

Vol.

90

Henning Bergenholtz, Sandro Nielsen & Sven Tarp (eds) Lexicography at a Crossroads. Dictionaries and Encyclopedias Today, Lexicographical Tools Tomorrow. 372 pages. 2009. ISBN 978-3-03911-799-4

Vol.

91

Manouchehr Moshtagh Khorasani The Development of Controversies. From the Early Modern Period to Online Discussion Forums. 317 pages. 2009. ISBN 978-3-3911-711-6

Vol.

92

María Luisa Carrió-Pastor (ed.) Content and Language Integrated Learning. Cultural Diversity. 178 pages. 2009. ISBN 978-3-3911-818-2

Vol.

93

Roger Berry Terminology in English Language Teaching. Nature and Use. 262 pages. 2010. ISBN 978-3-0343-0013-1

Vol.

94

Roberto Cagliero & Jennifer Jenkins (eds) Discourses, Communities, and Global Englishes 240 pages. 2010. ISBN 978-3-0343-0012-4

Vol.

95

Facchinetti Roberta, Crystal David, Seidlhofer Barbara (eds) From International to Local English – And Back Again. 268 pages. 2010. ISBN 978-3-0343-0011-7

Vol.

96

Cesare Gagliardi & Alan Maley (eds) EIL, ELF, Global English. Teaching and Learning Issues 376 pages. 2010. ISBN 978-3-0343-0010-0

Vol.

97

Sylvie Hancil (ed.) The Role of Prosody in Affective Speech. 403 pages. 2009. ISBN 978-3-03911-696-6

Vol.

98

Marina Dossena & Roger Lass (eds) Studies in English and European Historical Dialectology. 257 pages. 2009. ISBN 978-3-0343-0024-7

Vol.

99

Christine Béal Les interactions quotidiennes en français et en anglais. De l’approche comparative à l’analyse des situations interculturelles. 424 pages. 2010. ISBN 978-3-0343-0027-8

Vol. 100

Maurizio Gotti (ed.) Commonality and Individuality in Academic Discourse. 398 pages. 2009. ISBN 978-3-0343-0023-0

Vol. 101

Javier E. Díaz Vera & Rosario Caballero (eds) Textual Healing. Studies in Medieval English Medical, Scientific and Technical Texts. 213 pages. 2009. ISBN 978-3-03911-822-9

Notes on Contributors

Vol. 102

Nuria Edo Marzá The Specialised Lexicographical Approach. A Step further in Dictionary-making. 316 pages. 2009. ISBN 978-3-0343-0043-8

Vol. 103

Carlos Prado-Alonso, Lidia Gómez-García, Iria Pastor-Gómez & David Tizón-Couto (eds) New Trends and Methodologies in Applied English Language Research. Diachronic, Diatopic and Contrastive Studies. 348 pages. 2009. ISBN 978-3-0343-0046-9

Vol. 104

Françoise Salager-Meyer & Beverly A. Lewin Crossed Words. Criticism in Scholarly Writing? 371 pages. 2011. ISBN 978-3-0343-0049-0.

Vol. 105

Javier Ruano-García Early Modern Northern English Lexis. A Literary Corpus-Based Study. 611 pages. 2010. ISBN 978-3-0343-0058-2

Vol. 106

Rafael Monroy-Casas Systems for the Phonetic Transcription of English. Theory and Texts. 280 pages. 2011. ISBN 978-3-0343-0059-9

Vol. 107

Nicola T. Owtram The Pragmatics of Academic Writing. A Relevance Approach to the Analysis of Research Article Introductions. 311 pages. 2009. ISBN 978-3-0343-0060-5

Vol. 108

Yolanda Ruiz de Zarobe, Juan Manuel Sierra & Francisco Gallardo del Puerto (eds) Content and Foreign Language Integrated Learning. Contributions to Multilingualism in European Contexts 343 pages. 2011. ISBN 978-3-0343-0074-2

Vol. 109

Ángeles Linde López & Rosalía Crespo Jiménez (eds) Professional English in the European context. The EHEA challenge. 374 pages. 2010. ISBN 978-3-0343-0088-9

Vol. 110

Rosalía Rodríguez-Vázquez The Rhythm of Speech, Verse and Vocal Music. A New Theory. 394 pages. 2010. ISBN 978-3-0343-0309-5

Vol. 111

Anastasios Tsangalidis & Roberta Facchinetti (eds) Studies on English Modality. In Honour of Frank Palmer. 392 pages. 2009. ISBN 978-3-0343-0310-1

Vol. 112

Jing Huang Autonomy, Agency and Identity in Foreign Language Learning and Teaching. 400 pages. 2013. ISBN 978-3-0343-0370-5

Vol. 113

Mihhail Lotman & Maria-Kristiina Lotman (eds) Frontiers in Comparative Prosody. In memoriam: Mikhail Gasparov. 426 pages. 2011. ISBN 978-3-0343-0373-6

Vol. 114

Merja Kytö, John Scahill & Harumi Tanabe (eds) Language Change and Variation from Old English to Late Modern English. A Festschrift for Minoji Akimoto 422 pages. 2010. ISBN 978-3-0343-0372-9

Vol. 115

Giuliana Garzone & Paola Catenaccio (eds) Identities across Media and Modes. Discursive Perspectives. 379 pages. 2009. ISBN 978-3-0343-0386-6

399

400

Notes on Contributors

Vol. 116

Elena Landone Los marcadores del discurso y cortesía verbal en español. 390 pages. 2010. ISBN 978-3-0343-0413-9

Vol. 117

Maurizio Gotti & Christopher Williams (eds) Legal Discourse across Languages and Cultures. 339 pages. 2010. ISBN 978-3-0343-0425-2

Vol. 118

David Hirsh Academic Vocabulary in Context. 217 pages. 2010. ISBN 978-3-0343-0426-9

Vol. 119

Yvonne Dröschel Lingua Franca English. The Role of Simplification and Transfer. 358 pages. 2011. ISBN 978-3-0343-0432-0

Vol. 120

Tengku Sepora Tengku Mahadi, Helia Vaezian & Mahmoud Akbari Corpora in Translation. A Practical Guide. 135 pages. 2010. ISBN 978-3-0343-0434-4

Vol. 121

Davide Simone Giannoni & Celina Frade (eds) Researching Language and the Law. Textual Features and Translation Issues. 278 pages. 2010. ISBN 978-3-0343-0443-6

Vol. 122

Daniel Madrid & Stephen Hughes (eds) Studies in Bilingual Education. 472 pages. 2011. ISBN 978-3-0343-0474-0

Vol. 123

Vijay K. Bhatia, Christopher N. Candlin & Maurizio Gotti (eds) The Discourses of Dispute Resolution. 290 pages. 2010. ISBN 978-3-0343-0476-4

Vol. 124

Davide Simone Giannoni Mapping Academic Values in the Disciplines. A Corpus-Based Approach. 288 pages. 2010. ISBN 978-3-0343-0488-7

Vol. 125

Giuliana Garzone & James Archibald (eds) Discourse, Identities and Roles in Specialized Communication. 419 pages. 2010. ISBN 978-3-0343-0494-8

Vol. 126

Iria Pastor-Gómez The Status and Development of N+N Sequences in Contemporary English Noun Phrases. 216 pages. 2011. ISBN 978-3-0343-0534-1

Vol. 127

Carlos Prado-Alonso Full-verb Inversion in Written and Spoken English. 261 pages. 2011. ISBN 978-3-0343-0535-8

Vol. 128

Tony Harris & María Moreno Jaén (eds) Corpus Linguistics in Language Teaching. 214 pages. 2010. ISBN 978-3-0343-0524-2

Vol. 129

Tetsuji Oda & Hiroyuki Eto (eds) Multiple Perspectives on English Philology and History of Linguistics. A Festschrift for Shoichi Watanabe on his 80th Birthday. 378 pages. 2010. ISBN 978-3-0343-0480-1

Notes on Contributors

Vol. 130

Luisa Chierichetti & Giovanni Garofalo (eds) Lengua y Derecho. líneas de investigación interdisciplinaria. 283 pages. 2010. 978-3-0343-0463-4

Vol. 131

Paola Evangelisti Allori & Giuliana Garzone (eds) Discourse, Identities and Genres in Corporate Communication. Sponsorship, Advertising and Organizational Communication. 324 pages. 2011. 978-3-0343-0591-4

Vol. 132

Leyre Ruiz de Zarobe & Yolanda Ruiz de Zarobe (eds) Speech Acts and Politeness across Languages and Cultures. 402 pages. 2012. 978-3-0343-0611-9

Vol. 133

Thomas Christiansen Cohesion. A Discourse Perspective. 387 pages. 2011. 978-3-0343-0619-5

Vol. 134

Giuliana Garzone & Maurizio Gotti Discourse, Communication and the Enterprise. Genres and Trends. 451 pages. 2011. ISBN 978-3-0343-0620-1

Vol. 135

Zsuzsa Hoffmann Ways of the World’s Words. Language Contact in the Age of Globalization. 334 pages 2011. ISBN 978-3-0343-0673-7

Vol. 136

Cecilia Varcasia (ed.) Becoming Multilingual. Language Learning and Language Policy between Attitudes and Identities. 213 pages. 2011. ISBN 978-3-0343-0687-5

Vol. 137

Susy Macqueen The Emergence of Patterns in Second Language Writing. A Sociocognitive Exploration of Lexical Trails. 325 pages. 2012. ISBN 978-3-0343-1010-9

Vol. 138

Maria Vittoria Calvi & Giovanna Mapelli (eds) La lengua del turismo. Géneros discursivos y terminología. 365 pages. 2011. ISBN 978-3-0343-1011-6

Vol. 139

Ken Lau Learning to Become a Professional in a Textually-Mediated World. A Text-Oriented Study of Placement Practices. 261 pages. 2012. ISBN 978-3-0343-1016-1

Vol. 140

Sandra Campagna, Giuliana Garzone, Cornelia Ilie & Elizabeth Rowley-Jolivet (eds) Evolving Genres in Web-mediated Communication. 337 pages. 2012. ISBN 978-3-0343-1013-0

Vol. 141

Edith Esch & Martin Solly (eds) The Sociolinguistics of Language Education in International Contexts. 263 pages. 2012. ISBN 978-3-0343-1009-3

Vol. 142

Forthcoming.

Vol. 143

David Tizón-Couto Left Dislocation in English. A Functional-Discoursal Approach. 416 pages. 2012. ISBN 978-3-0343-1037-6

401

402

Notes on Contributors

Vol. 144

Margrethe Petersen & Jan Engberg (eds) Current Trends in LSP Research. Aims and Methods. 323 pages. 2011. ISBN 978-3-0343-1054-3

Vol. 145

David Tizón-Couto, Beatriz Tizón-Couto, Iria Pastor-Gómez & Paula Rodríguez-Puente (eds) New Trends and Methodologies in Applied English Language Research II. Studies in Language Variation, Meaning and Learning. 283 pages. 2012. ISBN 978-3-0343-1061-1

Vol. 146

Rita Salvi & Hiromasa Tanaka (eds) Intercultural Interactions in Business and Management. 306 pages. 2011. ISBN 978-3-0343-1039-0

Vol. 147

Francesco Straniero Sergio & Caterina Falbo (eds) Breaking Ground in Corpus-based Interpreting Studies. 254 pages. 2012. ISBN 978-3-0343-1071-0

Vol. 148

Forthcoming.

Vol. 149

Vijay K. Bhatia & Paola Evangelisti Allori (eds) Discourse and Identity in the Professions. Legal, Corporate and Institutional Citizenship. 352 pages. 2011. ISBN 978-3-0343-1079-6

Vol. 150

Maurizio Gotti (ed.) Academic Identity Traits. A Corpus-Based Investigation. 363 pages. 2012. ISBN 978-3-0343-1141-0

Vol. 151

Priscilla Heynderickx, Sylvain Dieltjens, Geert Jacobs, Paul Gillaerts & Elizabeth de Groot (eds) The Language Factor in International Business. New Perspectives on Research, Teaching and Practice. 320 pages. 2012. ISBN 978-3-0343-1090-1

Vol. 152

Paul Gillaerts, Elizabeth de Groot, Sylvain Dieltjens, Priscilla Heynderickx & Geert Jacobs (eds) Researching Discourse in Business Genres. Cases and Corpora. 215 pages. 2012. ISBN 978-3-0343-1092-5

Vol. 153

Yongyan Zheng Dynamic Vocabulary Development in a Foreign Language. 262 pages. 2012. ISBN 978-3-0343-1106-9

Vol. 154

Carmen Argondizzo (ed.) Creativity and Innovation in Language Education. 357 pages. 2012. ISBN 978-3-0343-1080-2

Vol. 155

David Hirsh (ed.) Current Perspectives in Second Language Vocabulary Research. 180 pages. 2012. ISBN 978-3-0343-1108-3

Vol. 156

Seiji Shinkawa Unhistorical Gender Assignment in Lahamon’s Brut. A Case Study of a Late Stage in the Development of Grammatical Gender toward its Ultimate Loss. 186 pages. 2012. ISBN 978-3-0343-1124-3

Vol. 157

Yeonkwon Jung Basics of Organizational Writing: A Critical Reading Approach. 151 pages. 2014. ISBN 978-3-0343-1137-3.

Notes on Contributors

403

Vol. 158

Bárbara Eizaga Rebollar (ed.) Studies in Linguistics and Cognition. 301 pages. 2012. ISBN 978-3-0343-1138-0

Vol. 159

Giuliana Garzone, Paola Catenaccio, Chiara Degano (eds) Genre Change in the Contemporary World. Short-term Diachronic Perspectives. 329 pages. 2012. ISBN 978-3-0343-1214-1

Vol. 160

Carol Berkenkotter, Vijay K. Bhatia & Maurizio Gotti (eds) Insights into Academic Genres. 468 pages. 2012. ISBN 978-3-0343-1211-0

Vol. 161

Beatriz Tizón-Couto Clausal Complements in Native and Learner Spoken English. A corpus-based study with Lindsei and Vicolse. 357 pages. 2013. ISBN 978-3-0343-1184-7

Vol. 162

Patrizia Anesa Jury Trials and the Popularization of Legal Language. A Discourse Analytical Approach. 247 pages. 2012. ISBN 978-3-0343-1231-8

Vol. 163

David Hirsh Endangered Languages, Knowledge Systems and Belief Systems. 153 pages. 2013. ISBN 978-3-0343-1232-5

Vol. 164

Eugenia Sainz (ed.) De la estructura de la frase al tejido del discurso. Estudios contrastivos español/italiano. 305 pages. 2014. ISBN 978-3-0343-1253-0

Vol. 165

Julia Bamford, Franca Poppi & Davide Mazzi (eds) Space, Place and the Discursive Construction of Identity. 367 pages. 2014. ISBN 978-3-0343-1249-3

Vol. 166

Rita Salvi & Janet Bowker (eds) Space, Time and the Construction of Identity. Discursive Indexicality in Cultural, Institutional and Professional Fields. 324 pages. 2013. ISBN 978-3-0343-1254-7

Vol. 167

Shunji Yamazaki & Robert Sigley (eds) Approaching Language Variation through Corpora. A Festschrift in Honour of Toshio Saito. 421 pages. 2013. ISBN 978-3-0343-1264-6

Vol. 168

Franca Poppi Global Interactions in English as a Lingua Franca. How written communication is changing under the influence of electronic media and new contexts of use. 249 pages. 2012. ISBN 978-3-0343-1276-9

Vol. 169

Miguel A. Aijón Oliva & María José Serrano Style in syntax. Investigating variation in Spanish pronoun subjects. 239 pages. 2013. ISBN 978-3-0343-1244-8

Vol. 170

Inés Olza, Óscar Loureda & Manuel Casado-Velarde (eds) Language Use in the Public Sphere. Methodological Perspectives and Empirical Applications 564 pages. 2014. ISBN 978-3-0343-1286-8

Vol. 171

Aleksandra Matulewska Legilinguistic Translatology. A Parametric Approach to Legal Translation. 279 pages. 2013. ISBN 978-3-0343-1287-5

404

Notes on Contributors

Vol. 172

Maurizio Gotti & Carmen Sancho Guinda (eds) Narratives in Academic and Professional Genres. 513 pages. 2013. ISBN 978-3-0343-1371-1

Vol. 173

Madalina Chitez Learner corpus profiles. The case of Romanian Learner English. 244 pages. 2014. ISBN 978-3-0343-1410-7

Vol. 174

Chihiro Inoue Task Equivalence in Speaking Tests. 251 pages. 2013. ISBN 978-3-0343-1417-6

Vol. 175

Gabriel Quiroz & Pedro Patiño (eds.) LSP in Colombia: advances and challenges. 339 pages. 2014. ISBN 978-3-0343-1434-3

Vol. 176

Catherine Resche Economic Terms and Beyond: Capitalising on the Wealth of Notions. How Researchers in Specialised Varieties of English Can Benefit from Focusing on Terms. 332 pages. 2013. ISBN 978-3-0343-1435-0

Vol. 177

Wei Wang Media representation of migrant workers in China. Identities and stances 198 pages. 2018. 978-3-0343-1436-7

Vol. 178

Cécile Desoutter & Caroline Mellet (dir.) Le discours rapporté: approches linguistiques et perspectives didactiques. 270 pages. 2013. ISBN 978-3-0343-1292-9

Vol. 179

Ana Díaz-Negrillo & Francisco Javier Díaz-Pérez (eds) Specialisation and Variation in Language Corpora. 341 pages. 2014. ISBN 978-3-0343-1316-2

Vol. 180

Pilar Alonso A Multi-dimensional Approach to Discourse Coherence. From Standardness to Creativity. 247 pages. 2014. ISBN 978-3-0343-1325-4

Vol. 181

Alejandro Alcaraz-Sintes & Salvador Valera-Hernández (eds) Diachrony and Synchrony in English Corpus Linguistics. 393 pages. 2014. ISBN 978-3-0343-1326-1

Vol. 182

Runhan Zhang Investigating Linguistic Knowledge of a Second Language. 207 pages. 2015. ISBN 978-3-0343-1330-8

Vol. 183

Hajar Abdul Rahim & Shakila Abdul Manan (eds.) English in Malaysia. Postcolonial and Beyond. 267 pages. 2014. ISBN 978-3-0343-1341-4

Vol. 184

Virginie Fasel Lauzon Comprendre et apprendre dans l’interaction. Les séquences d’explication en classe de français langue seconde. 292 pages. 2014. ISBN 978-3-0343-1451-0

Vol. 185

Forthcoming.

Vol. 186

Wei Ren L2 Pragmatic Development in Study Abroad Contexts 256 pages. 2015. ISBN 978-3-0343-1358-2

Vol. 187

Marina Bondi & Rosa Lorés Sanz (eds) Abstracts in Academic Discourse. Variation and Change. 361 pages. 2014. ISBN 978-3-0343-1483-1

Notes on Contributors

405

Vol. 188

Giuditta Caliendo Rethinking Community. Discourse, Identity and Citizenship in the European Union. 240 pages. 2017. ISBN 978-3-0343-1561-6

Vol. 189

Paola Evangelisti Allori (ed.) Identities in and across Cultures. 315 pages. 2014. ISBN 978-3-0343-1458-9

Vol. 190

Erik Castello, Katherine Ackerley & Francesca Coccetta (eds). Studies in Learner Corpus Linguistics. Research and Applications for Foreign Language Teaching and Assessment. 358 pages. 2015. ISBN 978-3-0343-1506-7

Vol. 191

Ruth Breeze, Maurizio Gotti & Carmen Sancho Guinda (eds) Interpersonality in Legal Genres. 389 pages. 2014. ISBN 978-3-0343-1524-1

Vol. 192

Paola Evangelisti Allori, John Bateman & Vijay K. Bhatia (eds) Evolution in Genre. Emergence, Variation, Multimodality. 364 pages. 2014. ISBN 978-3-0343-1533-3

Vol. 193

Jiyeon Kook Agency in Arzt-Patient-Gesprächen. Zur interaktionistischen Konzeptualisierung von Agency 271 pages. 2015. ISBN 978-3-0343-1666-8

Vol. 194

Susana Nicolás Román & Juan José Torres Núñez (eds) Drama and CLIL. A new challenge for the teaching approaches in bilingual education. 170 pages. 2015. ISBN 978-3-0343-1629-3

Vol. 195

Alessandra Molino & Serenella Zanotti (eds) Observing Norm, Observing Usage. Lexis in Dictionaries and in the Media. 430 pages. 2015. ISBN 978-3-0343-1584-5

Vol. 196

Begoña Soneira A Lexical Description of English for Architecture. A Corpus-based Approach. 267 pages. 2015. ISBN 978-3-0343-1602-6

Vol. 197

M Luisa Roca-Varela False Friends in Learner Corpora. A corpus-based study of English false friends in the written and spoken production of Spanish learners. 348 pages. 2015. ISBN 978-3-0343-1620-0

Vol. 198

Rahma Al-Mahrooqi & Christopher Denman Bridging the Gap between Education and Employment. English Language Instruction in EFL Contexts. 416 pages. 2015. ISBN 978-3-0343-1681-1

Vol. 199

Rita Salvi & Janet Bowker (eds) The Dissemination of Contemporary Knowledge in English. Genres, discourse strategies and professional practices. 171 pages. 2015. ISBN 978-3-0343-1679-8

Vol. 200

Maurizio Gotti & Davide S. Giannoni (eds) Corpus Analysis for Descriptive and Pedagogical Purposes. ESP Perspectives. 432 pages. 2014. ISBN 978-3-0343-1516-6

Vol. 201

Ida Ruffolo The Perception of Nature in Travel Promotion Texts. A Corpus-based Discourse Analysis. 148 pages. 2015. ISBN 978-3-0343-1521-0

406

Notes on Contributors

Vol. 202

Ives Trevian English suffixes. Stress-assignment properties, productivity, selection and combinatorial processes. 471 pages. 2015. ISBN 978-3-0343-1576-0

Vol. 203

Maurizio Gotti, Stefania Maci & Michele Sala (eds) Insights into Medical Communication. 422 pages. 2015. ISBN 978-3-0343-1694-1

Vol. 204

Carmen Argondizzo (ed.) European Projects in University Language Centres. Creativity, Dynamics, Best Practice. 371 pages. 2015. ISBN 978-3-0343-1696-5

Vol. 205

Aura Luz Duffé Montalván (ed.) Estudios sobre el léxico. Puntos y contrapuntos. 502 pages. 2016. ISBN 978-3-0343-2011-5

Vol. 206

Maria Pavesi, Maicol Formentelli & Elisa Ghia (eds) The Languages of Dubbing. Mainstream Audiovisual Translation in Italy. 275 pages. 2014. ISBN 978-3-0343-1646-0

Vol. 207

Ruth Breeze & Inés Olza (eds) Evaluation in media discourse. European perspectives. 268 pages. 2017. ISBN 978-3-0343-2014-6

Vol. 208

Vijay K. Bhatia & Maurizio Gotti (eds) Arbitration Discourse in Asia. 331 pages. 2015. ISBN 978-3-0343-2032-0

Vol. 209

Sofía Bemposta-Rivas, Carla Bouzada-Jabois, Yolanda Fernández-Pena, Tamara Bouso, Yolanda J. Calvo-Benzies, Iván Tamaredo (eds) New trends and methodologies in applied English language research III. Synchronic and diachronic studies on discourse, lexis and grammar processing. 280 pages. 2017. ISBN 978-3-0343-2039-9

Vol. 210

Francisco Alonso Almeida, Laura Cruz García & Víctor González Ruiz (eds) Corpus-based studies on language varieties. 285 pages. 2016. ISBN 978-3-0343-2044-3

Vol. 211

Juan Pedro Rica Peromingo Aspectos lingüísticos y técnicos de la traducción audiovisual (TAV). 177 pages. 2016. ISBN 978-3-0343-2055-9

Vol. 212

Maria Vender Disentangling Dyslexia. VenderPhonological and Processing Deficit in Developmental Dyslexia. 338 pages. 2017. ISBN 978-3-0343-2064-1

Vol. 213

Zhilong Xie Bilingual Advantages. Contributions of Different Bilingual Experiences to Cognitive Control Differences Among Young-adult Bilinguals. 221 pages. 2016. ISBN 978-3-0343-2081-8

Vol. 214

Larissa D’Angelo Academic posters. A textual and visual metadiscourse analysis. 367 pages. 2016. ISBN 978-3-0343-2083-2

Vol. 215

Evelyne Berger Prendre la parole en L2. Regard sur la compétence d’interaction en classe de langue. 246 pages. 2016. ISBN 978-3-0343-2084-9

Notes on Contributors

407

Vol. 216

David Lasagabaster and Aintzane Doiz (eds) CLIL experiences in secondary and tertiary education: In search of good practices. 262 pages. 2016. ISBN 978-3-0343-2104-4

Vol. 217

Elena Kkese Identifying Plosives in L2 English: The Case of L1 Cypriot Greek Speakers. 317 pages. 2016. ISBN 978-3-0343-2060-3

Vol. 218

Sandra Campagna, Elana Ochse, Virginia Pulcini & Martin Solly (eds) Languaging in and across Communities: New Voices, New Identities. Studies in Honour of Giuseppina Cortese. 507 pages. 2016. ISBN 978-3-0343-2073-3

Vol. 219

Adriana Orlandi & Laura Giacomini (ed.) Defining collocation for lexicographic purposes. From linguistic theory to lexicographic practice. 328 pages. 2016. ISBN 978-3-0343-2054-2

Vol. 220

Pietro Luigi Iaia Analysing English as a Lingua Franca in Video Games. Linguistic Features, Experiential and Functional Dimensions of Online and Scripted Interactions. 139 pages. 2016. ISBN 978-3-0343-2138-9

Vol. 221

Dimitrinka G. Níkleva (ed.) La formación de los docentes de español para inmigrantes en distintos contextos educativos. 390 pages. 2017. ISBN 978-3-0343-2135-8

Vol. 222

Katherine Ackerley, Marta Guarda & Francesca Helm (eds) Sharing Perspectives on English-Medium Instruction. 308 pages. 2017. ISBN 978-3-0343-2537-0

Vol. 223

Juana I. Marín-Arrese, Julia Lavid-López, Marta Carretero, Elena Domínguez Romero, Ma Victoria Martín de la Rosa & María Pérez Blanco (eds) Evidentiality and Modality in European Languages. Discourse-pragmatic perspectives. 427 pages. 2017. ISBN 978-3-0343-2437-3

Vol. 224

Gilles Col Construction du sens : un modèle instructionnel pour la sémantique. 292 pages. 2017. ISBN 978-3-0343-2572-1

Vol. 225

Ana Chiquito & Gabriel Quiroz (eds) Pobreza, Lenguaje y Medios en América Latina. 362 pages. 2017. ISBN 978-3-0343-2142-6

Vol. 226

Xu Zhang English Quasi-Numeral Classifiers. A Corpus-Based Cognitive-Typological Study. 360 pages. 2017. ISBN 978-3-0343-2818-0

Vol. 227

María Ángeles Orts, Ruth Breeze & Maurizio Gotti (eds) Power, Persuasion and Manipulation in Specialised Genres. Providing Keys to the Rhetoric of Professional Communities. 368 pages. 2017. ISBN 978-3-0343-3010-7

Vol. 228

Maurizio Gotti, Stefania Maci & Michele Sala (eds) Ways of Seeing, Ways of Being: Representing the Voices of Tourism. 453 pages. 2017. ISBN 978-3-0343-3031-2

Vol. 229

Dino Selvaggi Plurilingual Code-Switching between Standard and Local Varieties. A Socio-Psycholinguistic Approach 371 pages. 2018. ISBN 978-3-0343-2663-6

408

Notes on Contributors

Vol. 230

Anca-Cristina Sterie Interprofessional interactions at the hospital. Nurses’ requests and reports of problems in calls with physicians. 371 pages. 2017. ISBN 978-3-0343-2734-3

Vol. 231

Xiaodong Zhang Understanding Chinese EFL Teachers’ Beliefs and Practices in the Textbook-Based Classroom. 189 pages. 2017. ISBN 978-3-0343-3053-4

Vol. 232

Manuela Caterina Moroni & Federica Ricci Garotti (Hrsg.) Brücken schlagen zwischen Sprachwissenschaft und DaF-Didaktik. 345 pages. 2017. ISBN 978-3-0343-2667-4

Vol. 233

Dimitrinka Georgieva Níkleva Necesidades y tendencias en la formación del profesorado de español como lengua extranjera 401 pages. 2017. ISBN 978-3-0343-2946-0

Vol. 234

Juan Santana-Lario & Salvador Valera (Hrsg.) Competing patterns in English affixation. 272 pages. 2017. ISBN 978-3-0343-2701-5

Vol. 235

Francisco Salgado-Robles Desarrollo de la competencia sociolingüística por aprendices de español en un contexto de inmersión en el extranjero 241 pages. 2018. ISBN 978-3-0343-2323-9

Vol. 236

Maria Chiara Janner Sguardi linguistici sulla marca. Analisi morfosintattica dei nomi commerciali in italiano 345 pages. 2017. ISBN 978-3-0343-2667-4

Vol. 237

Bárbara Herrero Muñoz-Cobo & Otman El Azami Zalachi La primavera del árabe marroquí. 192 pages. 2017. ISBN 978-3-0343-3104-3

Vol. 238

Consuelo Pascual Escagedo El papel del oyente en la construcción de la conversación espontánea de estudiantes italianos en su interlengua y en su lengua materna 295 pages. 2017. ISBN 978-3-0343-3186-9

Vol. 239

Stefania M. Maci The MS Digby 133 Mary Magdalene. Beyond scribal practices: language, discourse, values and attitudes. 336 pages. 2017. ISBN 978-3-0343-3256-9

Vol. 240

Eliecer Crespo-Fernández Taboo in Discourse. Studies on Attenuation and Offence in Communication. 326 pages. 2018. ISBN 978-3-0343-3018-3

Vol. 241

Jana Altmanova, Maria Centrella, Katherine E. Russo (eds) Terminology & Discourse / Terminologie et discours. 424 pages. 2018. ISBN 978-3-0343-2417-5

Vol. 242

Xavier Blanco et Inès Sfar (dir.) Lexicologie(s) : approches croisées en sémantique lexicale. 442 pages. 2018. ISBN 978-3-0343-3056-5

Notes on Contributors

Vol. 243

Yunfeng Ge Resolution of Conflict of Interest in Chinese Civil Court Hearings. A Perspective of Discourse Information Theory. 302 pages. 2018. ISBN 978-3-0343-3313-9

Vol. 244

Carla Vergaro Illocutionary Shell Nouns in English 322 pages. 2018. ISBN 978-3-0343-3069-5

Vol. 245

Paolo Frassi L’adjectif en français et sa définition lexicographique. 270 pages. 2018. ISBN 978-3-0343-3394-8

Vol. 246

Suwilai Premsrirat and David Hirsh (eds) Language Revitalization. Insights from Thailand 328 pages. 2018. ISBN 978-3-0343-3497-6

Vol. 247

Wei Wang Researching Learning and Learners in Genre-based Academic Writing Instruction 282 pages. 2018. ISBN 978-3-0343-3297-2

Vol. 248

Isusi Alabarte, Alberto & Lahuerta Martínez, Ana Cristina (eds) La comprensión lectora de lengua extranjera Estudio de los factores de familiaridad, interés, género y métodos de evaluación 336 pages. 2018. ISBN 978-3-0343-3493-8

Vol. 249

Mercedes Eurrutia Cavero Approche didactique du langage techno-scientifique Terminologie et discours 374 pages. 2018. ISBN 978-3-0343-3512-6

Vol. 250

Aurora Ruiz Mezcua (ed.) Approaches to Telephone Interpretation Research, Innovation, Teaching and Transference 268 pages. 2018. ISBN 978-3-0343-3330-6

Vol. 251

Morini Massimiliano A Day in the News A Stylistic Analysis of Newsspeak 188 pages. 2018. ISBN 978-3-0343-3507-2

Vol. 252

Ignacio Guillén-Galve & Ignacio Vázquez-Orta (eds.) English as a Lingua Franca and Intercultural Communication Implications and Applications in the Field of English Language Teaching 414 pages. 2018. ISBN 978-3-0343-2763-3

Vol. 253

Bianca Del Villano Using the Devil with Courtesy Shakespeare and the Language of (Im)Politeness 216 pages. 2018. ISBN 978-3-0343-2315-4

Vol. 254

David Hirsh (ed.) Explorations in Second Language Vocabulary Research 252 pages. 2018. ISBN 978-3-0343-2940-8

409

410

Notes on Contributors

Vol. 255

Tania Baumann (ed.) Reiseführer - Sprach- und Kulturmittlung im Tourismus / Le guide turistiche - mediazione linguistica e culturale in ambito turistico 270 pages. 2018. ISBN 978-3-0343-3402-0

Vol. 256

Ariadna Sánchez-Hernández & Ana Herraiz-Martínez (eds) Learning second language pragmatics beyond traditional contexts 376 pages. 2018. ISBN 978-3-0343-3437-2

Vol. 257

Albert Bastardas-Boada, Emili Boix-Fuster, Rosa Maria Torrens (eds) Family Multilingualism in Medium-Sized Linguistic Communities 336 pages. 2019. ISBN 978-3-0343-2536-3

Vol. 258

Yuyang Cai Examining the Interaction among Components of English for Specific Purposes Ability in Reading. The Triple-Decker Model 296 pages. 2020. 978-3-0343-2913-2

Vol. 259

Catia Nannoni Participe présent et gérondif dans la presse française contemporaine 176 pages. 2019. ISBN 978-3-0343-3631-4

Vol. 260

Nieves Rodríguez Pérez & Bárbara Heinsch (eds.) Contextos multilingües. Mediadores interculturales, formación del profesorado de lenguas extranjeras 289 pages. 2019. ISBN 978-3-0343-3768-7

Vol. 261

Giuliana Elena Garzone, Mara Logaldo, Francesca Santulli (eds.) Investigating Conflict Discourses in the Periodical Press 244 pages. 2020. ISBN 978-3-0343-3668-0

Vol. 262

Laura Nadal Lingüística experimental y contraargumentación 233 pages. 2019. ISBN 978-3-0343-3791-5

Vol. 263

Claudia Claridge & Merja Kytö (eds.) Punctuation in Context – Past and Present Perspectives 288 pages. 2019. ISBN 978-3-0343-3790-8

Vol. 264

Maurizio Gotti, Stefania Maci, Michele Sala (eds.) Scholarly Pathways 530 pages. 2020. ISBN 978-3-0343-3860-8

Vol. 265

Ruth Breeze, Ana M. Fernández Vallejo (eds.) Politics and populism across modes and media 350 pages. 2020. ISBN 978-3-0343-3707-6

Vol. 266

Jean Marguerite Jimenez Understanding the Effects of Immediate Electronic Corrective Feedback on Second Language Development 252 pages. 2020. ISBN 978-3-0343-3815-8

Vol. 267

Sergio Rodríguez-Tapia, Adela González-Fernández (eds) Lenguas y turismo: estudios en torno al discurso, la didáctica y la divulgación 380 pages. 2020. ISBN 978-3-0343-3881-3

Notes on Contributors

411

Vol. 268

Ana Bocanegra-Valle (ed.) Applied Linguistics and Knowledge Transfer. Employability, Internationalisation and Social Challenges 344 pages. 2020. ISBN 978-3-0343-3714-4

Vol. 269

Beatrice Garzelli La traducción audiovisual español-italiano.Películas y cortos entre humor y habla soez 202 pages. 2020. ISBN 978-3-0343-4013-7

Vol. 270

Iván Tamaredo Complexity, Efficiency, and Language Contact. Pronoun Omission in World Englishes 292 pages. 2020. ISBN 978-3-0343-3902-5

Vol. 271

Silvia Domenica Zollo, Origine et histoire du vocabulaire des arts de la table. Analyse lexicale et exploitation de corpus textuels 239 pages. 2020. ISBN 978-3-0343-3890-5

Vol. 272

Paola Paissa, Michelangelo Conoscenti, Ruggero Druetta, Martin Solly (eds.) Metaphor and Conflict / Métaphore et conflit 385 pages. 2021. 978-3-0343-4068-7

Vol. 273

María Martínez-Atienza de Dios, Entre el léxico y la sintaxis: las fases de los eventos 178 pages. 2021. 978-3-0343-4173-8

Vol. 274

Noelia Castro-Chao, Argument Structure in Flux. The Development of Impersonal Constructions in Middle and Early Modern English, with Special Reference to Verbs of Desire. 300 pages. 2021. 978-3-0343-4189-9

Vol. 275

Gabriella Carobbio, Cécile Desoutter, Aurora Fragonara (eds.) Macht, Ratio und Emotion: Diskurse im digitalen Zeitalter / Pouvoir, raison et émotion: les discours à l’ère du numérique 246 pages. 2020. 978-3-0343-4184-4

Vol. 276

Miguel Fuster-Márquez, José Santaemilia, Carmen Gregori-Signes, Paula Rodríguez-Abruñeiras (eds.) Exploring discourse and ideology through corpora 293 pages. 2021. 978-3-0343-3969-8

Vol. 277

Tamara Bouso Changes in Argument Structure The Transitivizing Reaction Object Construction 392. 2021. 978-3-0343-4095-3

Vol. 278

Maria Luisa Maggioni, Amanda Murphy (eds.) Back to the Future. English from Past to Present 250 pages. 2021. 978-3-0343-4273-5

Vol. 279

Luisa Chierichetti Diálogos de serie. Una aproximación a la construcción discursiva de personajes basada en corpus 246 pages. 2021. 978-3-0343-4274-2

412

Notes on Contributors

Vol. 280

Cristina Lastres‐López From subordination to insubordination. A functional‐pragmatic approach to if/si‐constructions in English, French and Spanish spoken discourse 258 pages. 2021. 978-3-0343-4220-9

Vol. 281

Eleonora Federici, Stefania Maci (eds.) Gender issues. Translating and mediating languages, cultures and societies 500 pages. 2021. 978-3-0343-4022-9

Vol. 282

José Mateo, Francisco Yus (eds.) Metaphor in Economics and specialised Discourse 354 pages. 2021. 978-3-0343-4048-9

Vol. 283

Nicholas Brownlees (ed.) The Role of Context in the Production and Reception of Historical News Discourse 376 pages. 2021. 978-3-0343-4181-3

Vol. 284

Catalina Fuentes Rodriguez, María Ester Brenes Peña, Víctor Pérez Béjar (eds.) Sintaxis discursiva: construcciones y operadores en español 396 pages. 2021. 978-3-0343-4306-0

Vol. 285

Dominic Stewart Frequency in the dictionary. A corpus-assisted contrastive analysis of English and Italian 176 pages. 2021. 978-3-0343-4368-8

Vol. 286

Carla Bouzada-Jabois Nonfinite supplements in the recent history of English 320 pages. 2021. 978-3-0343-4226-1

Vol. 287

Sofía Bemposta Rivas Verb‐governed infinitival complementation in the recent history of English 320 pages. 2021. 978-3-0343-4227-8

Vol. 288

Mirella Agorni Translating Italy for the Nineteenth Century. Translators and an Imagined Nation in the Early Romantic Period 1816–1830s 180 pages. 2021. 978-3-0343-3612-3

Vol. 289

David Hirsh (ed.) Research Perspectives in Language and Education 276 pages. 2021. 978-3-0343-4219-3

Vol. 290

Federica Vezzani Terminologie numérique : conception, représentation et gestion 238 pages. 2022. 978-3-0343-4394-7

Vol. 291

Stefania M. Maci Evidential verbs in the genre of medical posters. A corpus-based analysis 498 pages. 2022. 978-3-0343-4521-7

Vol. 292

Francisco J. Álvarez-Gil Stance devices in tourism-related research articles: A corpus-based study. 170 pages. 2022. 978-3-0343-4555-2

Notes on Contributors

Vol. 293

Annalisa Baicchi / Stefania Biscetti (eds.) The Language of Fashion. Linguistic, Cognitive, and Cultural Insights 190 pages. 2022. 978-3-0343-4428-9

Vol. 294

Ana Maria Piquer-Píriz E-learning in EMI. Academic language for university students. 200 pages. 2022. 978-3-0343-4589-7

Vol. 295

Geneviève Henrot Sostero (ed.) Alle radici della fraseologia europea. 400 pages. 2023. 978-3-0343-4645-0

Vol. 296

Klara Dankova Les fibres textiles entre synchronie et diachronie: études terminologiques. 288 pages. 2023. 978-3-0343-4587-3

Vol. 297

Marta Carretero, Juana I. Marín-Arrese, Elena Domínguez Romero, Mª Victoria Martín de la Rosa (eds.) Evidentiality and Epistemic Modality. Conceptual and Descriptive Issues. 414 pages. 2023. 978-3-0343-3993-3

413

414

Notes on Contributors